Disruptive Technologies

ITUGET 2019 Background document
Emergency Telecommunications
Disruptive technologies
and their use in disaster risk
reduction and management
2019
International Telecommunication Union Disruptive technologies and their use in disaster risk reduction and management 2019
Telecommunication Development Bureau
Place des Nations
CH-1211 Geneva 20
Switzerland
Published in Switzerland
Geneva, 2019
Photo credits: Shutterstock
Disruptive technologies and
their use in disaster risk
The document was prepared by ITU expert Michael Minges, under the direction of the Least Developed
Countries, Small Island Developing States and Emergency Telecommunications (LSE) Division of the ITU
Telecommunication Development Bureau (BDT).
Please consider the environment before printing this report.
© ITU 2019
All rights reserved. No part of this publication may be reproduced, by any means whatsoever, without the prior
written permission of ITU.
Foreword
I am pleased to present this document on Disruptive technologies and their use in disaster risk
reduction and management, published to coincide with the third ITU Global Forum on Emergency
Telecommunications (GET-19), taking place from 6 to 8 March 2019 in Balaclava, Mauritius. The theme
of GET-19 – Innovating together to save lives: using technologies in disaster management – is highly
relevant given the devastating impacts that natural hazards have on people and economies across
the globe. Each year over the past decade there were, on average, 354 disasters, 68 000 deaths, 210
million people affected and USD 153 billion in damages. However, there are signs that loss of life from
major natural hazards are on the decline due to better disaster risk management.
This document discusses the use and opportunities of ICTs and disruptive technologies for disaster
risk reduction and management. It responds to requests from ITU Member States to identify relevant
technologies and facilitate the sharing of best practices. The document finds that technological
advancement and innovation are creating new opportunities for enhancing disaster resiliency and risk
reduction. Developments in disruptive technologies – such as artificial intelligence (AI), the Internet
of Things (IoT) and Big Data – and innovations in such areas as robotics and drone technology are
transforming many fields, including disaster risk reduction and management. The rapid spread of
supporting digital infrastructure and devices, especially wireless broadband networks, smartphones
and cloud computing, has created the foundation for the application of disruptive technologies for
disaster management. Disruptive technologies can spread critical information more quickly, improve
understanding of the causes of disasters, enhance early warning systems, assess damage in new ways
and add to the knowledge base of the social behaviours and economic impacts after a crisis strikes.
Situational awareness is improving with new tools providing the crisis community with a clearer
understanding of the extent of damage and where to prioritize resources.
At the same time, the pace, scope and impact vary among the technologies. Use of drones and IoT is
increasing, as experience is gained and costs fall. While social media are playing a greater role during
disasters and the public is using digital technologies such as crowdsourcing map details to support
disaster management, many uses of Big Data, robots and AI remain largely experimental. Large-
scale impacts will require more time and investments in skills and research. Traditional technologies,
though not considered disruptive, continue to play a critical role in disaster management, and are
also benefitting from digitization. Satellite imagery and seismometers remain important methods for
detecting, monitoring and accessing disasters, and text messaging has a wide reach when commu-
nicating with the public.
We set out important steps that governments, relief agencies, the private sector, the research
community and assistance agencies can take to maximize benefits from the opportunities identified
in this document, and we highlight the importance of regulation, training, scaling and building partner-
ships. These steps will help disruptive technologies achieve wider impact before, during, and after
disasters with the potential to significantly reduce loss of life and hasten recovery efforts.
iii
Executive Summary
Technological advancement and innovation have created new opportunities for enhancing disaster
resiliency and risk reduction. Developments in disruptive technologies – such as artificial intelligence
(AI), the Internet of Things (IoT), and Big Data – and innovations in such areas as robotics and drone
technology are transforming many fields, including disaster risk reduction and management. The
rapid spread of supporting digital infrastructure and devices – such as wireless broadband networks,
smartphones and cloud computing – has created the foundation for the application of disruptive
technologies for disaster management.
The first documented use of aerial drones was in 2005 after Hurricane Katrina in the United States
of America. Because roads were blocked by trees, drones were deployed to search for survivors and
assess river levels. Aerial drones are currently used for different disaster phases: preparedness, such
as monitoring volcanic activity in order to determine when warnings should be issued; response, such
as delivering equipment to locations where ground-based transportation has been disrupted; and
recovery, such as photographing disaster areas for damage assessments. Underwater drones were
used during the 2018 Hurricane Florence in the United States of America, measuring ocean heat
fuelling the hurricane and transmitting data to the National Weather Service. The data filled in gaps
left by satellite images, thus improving hurricane modelling. The data also enhanced forecasting the
intensity and route of the hurricane, and sensors attached to the drones measured salinity levels to
determine how much water from rain and rivers was mixing in the ocean.
An avalanche of data is being generated by sensors, closed-circuit television, smartphones, financial

transactions and Internet activities, to name just a few. While many of these data are being mined
by businesses for commercial purposes, Big Data analytics holds enormous potential for crisis
management. Examples include the analysis of social media communications during a disaster to
understand the types of data and creators, in order to have more impact and reduce false information.
Another example is the use of financial transactions to monitor economic activity during and after a
disaster, in order to improve targeting of support efforts.
The use of sensors for monitoring conditions that could trigger disasters dates back a number of years.
Improvements in cloud computing, broadband wireless networks, the sensors themselves and data
analysis have led to the emergence of powerful, integrated and real-time systems referred to as the
Internet of Things (IoT). Disaster management is an ideal use case for IoT applications, since sensors
can send alerts about a number of potentially dangerous situations. Tree sensors can detect if a fire
has broken out by testing temperature, moisture and carbon dioxide levels. Ground sensors can
detect earth movements, which might signal earthquakes. River levels can be monitored by sensors
for possible flooding.
AI could have a tremendous impact for disaster management regarding quickening recovery and
response times. Humanitarian groups are hoping to speed up map creation by using machine learning
to extract objects such as buildings and roads from aerial images. Considerable research is currently
being devoted to the use of AI for detecting and possibly one day predicting earthquakes.
Robots have become more sophisticated through integration with microprocessors and sensors. Their
growing dexterity makes them suitable for disaster situations that are too dangerous for humans or
rescue animals. Search-and-rescue robots were first used following the September 2011 terrorist
attack in New York City to assess the wreckage of the demolished World Trade Center. Since then,
more than 50 deployments of robots for disaster use have been reported. Breakthroughs are being
achieved in Japan, where there is the possibility for commercialization of robots designed specifically
for disasters.
iv
One challenge during a rapidly evolving disaster is coordinating and verifying information among
different stakeholders. The Blockchain distributed ledger system and chain of verified records could
play a significant role in improving information control. In the United States of America, the Centers
for Disease Control and Prevention is planning a pilot to test blockchain for more rapid and reliable
collection of data during a crisis in order to reduce the spread of disease. This has relevance for
disaster management since, similar to public health, agencies offering relief need to share trusted
data quickly to collaborate effectively.
Social media are playing a greater role during disasters. Twitter has been widely used by the relief
community in a number of disasters to coordinate response. Unlike text messages, users can follow
tweeters; ’cards‘ can be included with links to photos, videos and other media; and hashtags help
quickly receive or find tweets on a specific topic. The Facebook Crisis Response app allows users to
mark themselves as safe, reassuring friends and family, provide or seek help, donate money, and
receive information.
The public is using digital technologies to support disaster management. Crowdsourcing is helping
to add vital details to maps of disaster areas. This enhances satellite imagery by providing greater
granularity, making relief efforts more effective and targeted. It is also quicker and cheaper, especially
since the crowd often does this on a volunteer basis, compared with traditional methods. Digital
technologies provide another way of raising donations for disaster relief victims. More than USD 40
million was contributed via text messaging following the Haiti earthquake disaster in 2010. In India,
Twitter was used to alert people to crowdfunding sites for those affected by floods. Some relief
agencies have begun accepting donations in cryptocurrencies, such as Bitcoin. Mobile money is a
safe way for relief organizations to transfer funds to those affected by disasters.
Traditional technologies, though not considered disruptive, continue to play a critical role in disaster
management, and are also benefitting from digitization. Text messaging is used by many relief organi-
zations, because of the ubiquity of mobile phones. Satellite-based emergency mapping has improved
both in quality and response. New radar apertures can see through storm clouds, and satellite activa-
tions for disasters have risen. Radar radio waves continue to play a critical role for disaster monitoring,
for example high-frequency ocean radar systems used for monitoring tsunami activity.
The document takes a closer look at the application of particular disruptive technologies through
five case studies:
(a) In Vanuatu, aerial drones were used following Cyclone Pam for disaster assessment. Drones were
felt to be an ideal solution for rapid and granular evaluation of the situation, particularly since
cloud cover obscured satellite images. The imagery showed which houses were unrepairable
compared with those that could be fixed, helping to guide funding and recovery efforts. Crops
were surveyed for damage in order to determine how much food people would need from
other sources. The imagery was input to an open source mapping platform for volunteers to
upload and geo-tag images from social media to overlay on the map. There were challenges with
connectivity, the weather conditions and data formats. Nevertheless, the experience provided
valuable insights, with the drones providing the fastest method of mapping the damaged areas.
(b) Twitter was deployed by a number of different groups for a variety of purposes following severe
flooding in Chennai, India, in 2015. Although emergency telephone numbers were established,
they were overloaded, making social media such as Twitter a popular method for communication
for those with Internet access. The Twitter live, real-time public platform was used by non-
governmental organizations (NGOs), the public, government agencies and the media to share
and exchange information. This resulted in unprecedented collaboration on the platform in
responding to the crisis. Twitter India publicized three hashtags to be used during the flood,
depending on the nature of the tweet, for citizen groups to help agencies on the ground. A
variety of information was shared, including helpline phone numbers, updated train schedules,
weather forecasts, relief efforts and safety tips. This helped to magnify critical messages, organize
v
relief efforts, assist government agencies, warn the public and provide information in real time
to citizens trapped in the floods.
(c) Researchers used Big Data techniques to explore financial transactions before, during and after
Hurricane Odile struck the Mexican State of Baja California Sur in September 2014, to analyse
its economic impact on those affected. The analysis identified which groups were most affected
for targeting post-disaster assistance and how long it took to return to normal, and generated
estimates of the economic impact. It would be attractive to develop an ex ante model so that
the financial response to disasters could be analysed in real time to provide ongoing feedback
loops to relief efforts.
(d) Japan is a pioneering centre for the use of robots in disasters. The Human–Robot Informatics
Laboratory of Tohoku University has developed several types of robots for disaster response,
including a snake-like robot with a camera that can crawl over obstacles, follow walls, and make
turns in tight spaces. Research is ongoing to augment search-and-rescue dogs using cameras,
Global Positioning System (GPS) and inertial measurement units, which are being designed to
be small and light to fit on a dog pack. Damage to the Fukushima Daiichi Nuclear Power Plant in
the 2011 earthquake triggered significant robot research due to radiation preventing humans
from carrying out direct cleanup activities. One of the biggest challenges has been determining
what happened to the fuel inside the core of the reactor. Various robots were used to penetrate
the core, but without success. Finally, in 2017, a small robot designed to operate underwater
with severe radiation exposure succeeded in locating the missing fuel inside the reactor core.
Japanese manufacturer Honda is developing a disaster response robot that can walk, scale
obstacles and climb ladders. If the prototype reaches fruition, it could have a major impact for
rescue operations and clearing hazardous materials.
(e) Following the flooding of the Liboriana River, which triggered a devastating landslide in May
2015, causing more than 80 deaths, the Government of Colombia’s National Unit for Disaster
Risk Management took steps to mitigate future occurrences. It hired a company to implement
an early warning system using IoT technology. Five solar powered sensors were installed along
the Liboriana and two other rivers to monitor water levels and air temperature using ultrasound.
The use of solar ensures the sensors continue to function in the event of an electricity outage.
The system automatically sends a text message to village authorities if a risk is detected, and
data are also stored in the cloud for others to access.
The volume of seismic data has increased dramatically, providing fuel for solutions to detect and locate
earthquakes. Scientists have based detection on continuous seismic records, searching for repeating
signals that provide information on upcoming earthquakes. Analysing many years of earthquake
waveforms requires considerable computing power. A new approach uses AI to reduce waveform
analysis processing. Based on convolutional neural network computer learning, the software is trained
to analyse large waveform data sets, and distinguish between noise and true earthquake signals, while
preserving location information. Compared with other methods, the convolutional neural network
method is faster and retains location information with a high degree of accuracy.
These examples illustrate how disruptive technologies today are refining processes by spreading
critical information more quickly, improving understanding of the causes of disasters, enhancing early
warning systems, assessing damage quickly and adding to the knowledge base of the social behaviours
and economic impacts after a crisis strikes.
Application of disruptive technologies to disaster management vary in pace, scope and impact. Social
media platforms such as Facebook and Twitter have been applied in a number of events, and aerial
drones and IoT are increasing in use as experience is gained and costs fall. Older technologies such
as satellite imagery and seismometers are still the most important methods for detecting, monitoring
and accessing disasters, and text messaging has the widest reach for communicating with the public.
Big data, robots and AI remain largely experimental, and large-scale impacts will require more time
and investments in skills and research.
vi
Several recommendations have been identified that governments, relief agencies, the private sector
and assistance agencies can take to maximize benefits offered by disruptive technologies:
(a) Systemization and standardization are needed to improve the application of technology
interventions. Open standards will help to lower costs, ensure interoperability and enhance
scaling. The standardization should also extend to the use of Big Data, which is currently often
shrouded in opaqueness. Clear and transparent sharing protocols should be implemented,
including application programming interfaces. For social media, standardized hashtags should
be employed to reduce confusion among the public and magnify impacts.
(b) Reach of digital technologies must be factored into disaster management strategies. In respect
to communications among stakeholders, this includes considering the purpose and audience.
While Twitter has proven useful in crisis situations, particularly among the relief community, its
penetration is relatively low among the general public. It should also be considered that some
people may not want to use proprietary platforms for a number of reasons, and therefore relying
on only one method may not reach all intended recipients.
(c) A global repository featuring information on how digital technologies are being applied for
disaster management would raise awareness and understanding. Hundreds of applications
of disruptive technology are underway around the world, but experiences are often buried in
news articles and research reports. An information base would be useful for identifying digital
interventions that have worked, who the implementers were, and other material to increase
understanding about which technologies are relevant for different country circumstances and
types of disasters.
(d) Partnerships with the private sector and academia will be critical for understanding and applying
digital technologies for disaster prediction, detection, response and relief. Numerous uses of
disruptive technologies are being developed by the private sector. In addition, the private sector
controls significant amounts of personal information in Big Data sets, which are of immense use
for the disaster community. Similarly, considerable relevant research is being undertaken by the
academic community.
(e) Scaling disruptive technologies for crisis is essential to have widespread impact and lower
deployment costs. To date, many interventions are still pilots or carried out in an ad hoc informal
manner. Processes should exist to identify relevant use cases and scale them. Given the vast
potential of disruptive technologies for disaster management under a huge variety of different
circumstances, there is a need to nurture innovation. This is particularly relevant given the wide
country contexts around the world, different types of disasters and crisis phases. Examples from
the start-up world are relevant where incubators, labs, competitions, and venture capital are
used to discover, mentor and scale up promising innovations.
(f) Training is indispensable for the disaster community to understand how to properly and
responsibly deploy new and emerging digital technologies in crisis settings. Manuals are needed
for different technologies. For example, in the case of social media, this would cover hashtag
guidelines, usage by public and relief organizations, coping with fake information, etc. Exchanges
should be arranged for disaster management personnel to gain experience using new tools.
(g) Legal ramifications of technological research and interventions for disasters need to be
understood. This is fairly straightforward in respect to specific regulations, such as registration
and regulation of drones, but more nebulous regarding data protection and privacy. One
relevant dilemma is whether the lack of data protection and privacy laws inhibits or encourages
technologies that make heavy use of personal information. Further, while Big Data may eventually
be forthcoming, in the event of a disaster it is needed immediately, so it is essential to have data
access and sharing protocols in place beforehand. The disaster community has developed codes
of conduct in certain areas that can help when laws are vague.
(h) Adequate capacity remains fundamental for properly planning and deploying relevant digital
technologies. While digital technologies show great promise for all phases of disasters, on the
vii
ground, planning, management and operations are critical to their success. Disaster agencies do
not need to be experts in digital technologies, but they do need to understand enough about
them to develop proactive blueprints for deploying them. Disaster agencies might also consider
creating a chief technologist post to better understand how to apply disruptive technologies.
viii
Table of Contents
Foreword iii
Executive Summary iv
1 Background 1
2 Disruptive technologies for disaster risk reduction and management 3

2.1 Mobile phones 4
2.2 Drones 6
2.2.1 Air 6
2.2.2 Underwater 7
2.3 Big Data 8
2.4 Internet of Things 11
2.5 Artificial intelligence 12
2.6 Robots 13
2.7 Blockchain 13
2.8 Social media 13
2.8.1 Twitter 13
2.8.2 Facebook 15
2.9 Crowdsourcing and crowdfunding 17
2.10 Others 18
3 Case studies 20
3.1 Drones for disaster damage assessment in Vanuatu 20
3.2 Twitter during the Chennai flood, India 21
3.3 Financial big data for Hurricane Odile, Mexico 24
3.4 Disaster robots in Japan 25
3.5 Internet of Things for river flooding control in Colombia 27
3.6 Artificial intelligence for earthquake detection and prediction 28
4 Challenges, relevance and sustainability 32

4.1 Challenges 32
4.2 Relevance and sustainability 35
5 Conclusions and recommendations 37
Abbreviations 40
References 41
ix
List of Tables, Figures and Boxes
Tables
Table 2.1: Radio, television and mobile phone (% of households), selected LDCs 19
Table 3.1: Account ownership by income group, 2017 25
Table 3.2: Comparison of performance of earthquake detection methods 30
Table 4.1: Disruptive technology classification (based on case studies) 36
Figures
Figure 1.1: Annual disasters and economic impact 1
Figure 1.2: Disaster management phases 2
Figure 1.3: Natural disaster subgroup 3
Figure 2.1: Crowdsourcing roles based on user involvement and level of data processing 5
Figure 2.2: Indago UAV 7
Figure 2.3: Slocum glider UUV 8
Figure 2.4: Taxonomy of Big Data and crisis analytics 9
Figure 2.5: Global Twitter users 14
Figure 2.6: Facebook Safety Check 15
Figure 2.7: Number and distribution of Facebook users 16
Figure 2.8: 2G network coverage following volcanic eruption in Guatemala 16
Figure 2.9: Putting communities on a map 17
Figure 2.10: Number of disasters with satellite response 19
Figure 3.1: Mapbox of Cyclone Pam damage 21
Figure 3.2: Twitter India hashtags for the Chennai floods 22
Figure 3.3: Twitter link to crowdfunding for Chennai floods 23
Figure 3.4: Expected versus real banking transactions in Baja California Sur, September—
October 2014 24
Figure 3.5: Little sunfish robot 26
Figure 3.6: Honda E2-DR robot 27
Figure 3.7: IoT system for river monitoring in Salgar, Colombia 28
Figure 3.8: Seismogram plot 29
Figure 3.9: Earthquake sensor density in California and Japan 31
Boxes
Box 2.1: 5G and disruptive technologies 4
Box 2.2: United Nations Global Pulse: Big Data and Disaster Management initiatives and
tools10
Box 2.3: Artificial intelligence and machine learning in emergency situations 12
Box 2.4: Google Crisis Map 18
Box 3.1: Using ground sensor technology to detect earthquakes 31
Box 4.1: Humanitarian UAV Code of Conduct 33
x
Disruptive technologies and their use in disaster risk reduction and management
1 Background
Technological advancements and innovation create new possibilities for supporting disaster resiliency
and risk reduction actions. Developments in the area of AI, IoT, and fifth generation (5G) wireless
network, Big Data, and innovations in such areas as robotics and drone technology are transforming
many fields, including disaster risk reduction and management. IoT, Big Data and AI are key drivers
behind the ongoing digital transformation, and will play an increasingly important role in all phases
of disaster management and resiliency development. Examples include the use of AI to analyse
seismometer data to make detection models about earthquakes, and the use of Big Data to identify
communication patterns during disasters, through the analysis of social media.
Disasters have devastating effects on people’s lives, as well as significant economic impact. From
2007 to 2017, the yearly average of the impact was 354 disasters, 68 000 deaths, 210 million people
affected and USD 153 billion in damages (Centre for Research on the Epidemiology of Disasters (CRED),
2018). In 2017 alone, 122 countries were affected by disasters. There has been a marked increase
reported in the cost of damage caused by disasters in the ten years up to the year 2000, in addition
to a sustained increase in both the number and cost of disasters from 2000 to 2018 (Figure 1.1).
Figure 1.1: Annual disasters and economic impact
Source: EM-DAT: The Emergency Events Database – Université catholique de Louvain (UCL) – CRED, – www.emdat.be,
Brussels, Belgium.
The 2030 Agenda for Sustainable Development has helped raise awareness of the importance of
disaster and emergency management: of the 17 Sustainable Development Goals established by the
United Nations, four of them (Goals 1, 2, 11 and 13) refer to the need of nations and communities
to address the challenges related to natural hazards and disasters.
Member States have provided ITU with a clear mandate to ensure that all countries are able to take
advantage of the opportunities brought by information and communication technologies (ICTs),
including new technologies to address the challenges of disasters. The 2017 World Telecommunication
Development Conference reiterated the need for ITU to identify relevant technologies and facilitate
the sharing of best practices and applications, to allow countries to benefit from ICTs for disaster risk
reduction and management.1
1
ITU Resolution 34 (Rev. Buenos Aires, 2017). The role of telecommunications/information and communication
technology in disaster preparedness, early warning, rescue, mitigation, relief and response. Available from www.itu
.int/en/ITU-D/Conferences/WTDC/WTDC17/Documents/WTDC17_FinalReport_en.pdf (accessed 2 February 2019).
1
ITU has been dealing with the issue of emerging and disruptive technologies while designing and
implementing projects and organizing events, including the annual AI for Good Global Summit.2 The
ITU Telecommunication Development Sector (ITU-D) Study Group 2 has a Question on the topic
(Utilization of telecommunications/ICTs for disaster preparedness, mitigation and response) and ITU
works in this area through its ITU-D programme on emergency telecommunications.
Introduction, scope, and purpose
This document discusses the use and opportunities of ICTs and disruptive technologies for disaster
risk reduction and management. When the term disruptive technology was introduced more than
two decades ago, it referred to new technologies having an impact on incumbent businesses with
scope for major disruption (Christensen et al., 2015). Today, the term has taken on a wider lens by
including the innovative application of new technologies across a range of domains, including disaster
risk reduction and management.
The document looks at examples of the application of emerging technologies and their current and
potential impact for managing and reducing the impact of disasters. The document covers emerging
technologies and addresses existing and new solutions for disaster risk reduction and emergency
management, and informed decision-making linked to the four phases of disaster management
(Figure 1.2).
Figure 1.2: Disaster management phases
Source : ITU
The focus of this document is on disaster management in the case of natural hazards (geophysical,
hydrological, meteorological and climatological) (Figure 1.3). It is intended to serve as a reference
for ITU Members, the humanitarian community and others on the use and implementation of such
solutions. It also seeks to contribute to raising awareness in terms of opportunities and challenges
associated with the deployment and use of such technologies. The examples and case studies cover
different regions, including both developed and developing countries, and different hazards.
2
See https://aiforgood.itu.int/accessed 2 February 2019).
2
Figure 1.3: Natural disaster subgroup
Source: CRED, Institute of Health and Society, and Université catholique de Louvain, 2016.
Some of the issues covered by the document include the complexity of digital innovations and their
use in the context of disaster risk reduction and management in terms of impact, costs, policy and
regulation, and skills. It also identifies different stakeholders involved in the use of the technologies.
2 Disruptive technologies for disaster risk reduction and

management
This chapter offers an overview of disruptive technologies for disaster risk reduction and management.
It identifies and defines key technologies, including references to research and projects.
The rapid spread of digital infrastructure and devices has created immense potential for the use of
disruptive technologies for disaster management. Mobile broadband technologies are being rapidly
extended, with an estimated 90 per cent of the world’s population covered by at least a 3G signal by
the end of 2018. Smartphones are proliferating; they can capture the geographic location of the user
to help locate people affected by disasters. Furthermore, smartphones enable users to communicate
in a richer way than basic mobile devices do, and use applications such as social media to rapidly
exchange information during a crisis. Cloud computing enables storage of data generated by different
sources and sharing among different groups (e.g. users, governments and NGOs). The monitoring
sensors capture a variety of information, whether worn by users or embedded in the ground, providing
real-time data streams and forming the basis of the Internet of Things (IoT). The spread and availability
of these technologies vary among developed and developing nations and among high- and low-income
regions, and this digital divide influences their suitability for different disaster management scenarios.
Availability of the latest versions of technologies affects functionality and applicability. For example,
5G wireless networks are viewed as a key enabler of IoT,3 but their deployment will initially occur in
urban areas and predominantly in developed countries.
Classification of disruptive technologies varies. Some are universal and primarily concerned with the
flow and analysis of communicating information generated by citizens, governments and sensors
before, during and after a crisis (e.g. social media and Big Data). Others, such as drones and robots,
are hardware aimed at specific interventions (though they also transmit essential data for analysis).
Other digital technologies, while not necessarily considered disruptive, are critical, since they are
3
See Ericsson, 5G and IoT: Ushering in a new era, available from www.ericsson.com/en/about-us/india/authored
-articles/5g-and-iot-ushering-in-a-new-era (accessed 3 February 2019).
3
widely deployed and available, and arguably currently still have a greater impact (e.g. text messages
and satellite images). Some technologies can be integrated with others to magnify the impact. That
sometimes makes it difficult to clearly delineate them: for instance, AI is used in Big Data analytics as
well as robotics (which sometimes includes the category of drones).
Box 2.1: 5G and disruptive technologies

The 5G wireless technology is viewed as a key enabler of several disruptive technologies
applicable for disaster situations. IoT sensors generate vast amounts of data that need
to be communicated rapidly. Drones are more effective if high definition images can be
transmitted in real time, instead of having to wait until they return to their base. The 5G
technology has higher capacity, is faster and has lower latency compared with previous
generations, and thus can support disruptive technologies reaching their full functionality.
In early 2012, ITU embarked on a programme to develop IMT (International Mobile

Telecommunication) for 2020 and beyond, setting the stage for 5G research activities
that are emerging around the world. In September 2015, ITU finalized its vision of the 5G
mobile broadband-connected society. This view of the horizon for the future of mobile
technology will be instrumental in setting the agenda for the World Radiocommunication
Conference 2019, where deliberations on additional spectrum for 5G are taking place. The
whole process is planned to be completed in 2020, when detailed specifications for the
new radio interfaces, based on input from the 3rd Generation Partnership Project (3GPP)
and other stakeholders, will be submitted for approval within ITU. Meanwhile, other ITU
standardization activities are addressing 5G enabling technologies, including extremely
high-capacity optical networks that form the backbone of 5G (virtual networks, network
function virtualization and network slicing), as well as strategies to assist machine learning
in contributing to the efficiency of emerging 5G systems.
According to the Global Mobile Suppliers Association, almost a dozen countries had deployed
5G by January 2019, and operators in around 50 countries have announced plans to launch
5G before the end of 2022.
2.1 Mobile phones

As mobile phones have evolved in functionality, their impact for disaster relief has grown. From voice
calls to text messages – and now location-based services, cameras and Internet access – mobile
phones have a diverse set of features being leveraged by the public and disaster community in times
of crisis. The wide spread of mobile phones – often with a higher penetration than television or radio
in developing nations – today makes them the most universal communication device in the world.
Simple voice calls put the public in touch with friends, family and relief workers during times of crisis,
with immeasurable lifesaving impacts. However, voice transmission lines are often overloaded during
disasters. In this case, SMS (Short Message Service) text messaging, is more effective, since it uses
a control rather than a voice channel for transmission. Text messages are asynchronous so, unlike a
phone call, they are temporarily stored while waiting to be delivered.
SMS messages of up to 160 alphanumeric characters can be sent from one person to another or
broadcast to a group of targeted recipients (i.e. Cell Broadcast4). Unlike other message-like applica-
tions, which require a computer or smartphone and Internet access, text messaging is a standard
feature on all mobile phones, and thus its reach is immense. According to a 2016 study of mobile
4
Available from www.cellbroadcastforum.org/whatisCB/index.html (accessed 3 February 2019).
4
phone users in 31 countries, 74 per cent used text messaging, the most popular mobile phone activity
after making phone calls.5
SMS provides information quickly to those touched by disasters. Several studies have examined the
impact of text messaging during disasters. Meier and Munro (2010) discuss the use of SMS following
the major earthquake that struck Haiti in January 2010, destroying the capital Port-au-Prince. A
short SMS code – 4636 – was created for people to text their locations, with volunteers using the
information to create a live crisis map to support relief assistance. As the first widespread use of text
messaging, invaluable lessons were learned, such as making the service ’opt-in‘, targeting messages
to specific geographies and groups, setting up SMS platforms prior to the disaster, and minimizing
use of short codes. These lessons have been codified in an SMS Code of Conduct for natural disasters
(GSMA, 2013).
The advent of smartphones has created new opportunities for the public (i.e. crowd) to assist –
knowingly or unknowingly – in helping to respond to disasters. Four roles have been identified (Poblet
et al., 2014) (Figure 2.1):
(a) Crowd as sensors: Mobile phones are continuously generating data from their internal sensors,
including GPS, accelerometers, gyroscopes and magnetometers. The data are collected
(opportunistic crowdsourcing) with little, if any, data processing by the user.
(b) Crowd as social computers: Users generate data by using apps such as those for social media.
These data are collected by platforms (Big Data). Like the crowd as sensor, there is no direct
effort to share the data by the user.
(c) Crowd as reporters: Users offer their own information on events (e.g. taking a photo of damage,
tweeting about weather conditions, etc.). This user-generated content can include supplementary
information (e.g. hashtags).
(d) Crowd as microtaskers: Users create content such as adding roads or buildings to satellite images.
Here, users are active participants and often have specific skills.
Figure 2.1: Crowdsourcing roles based on user involvement and level of data processing
Source: Poblet et al. 2014.
5
Deloitte. 2016. Global mobile consumer trends. https://www2.deloitte.com/content/dam/Deloitte/global/Documents/
Technology-Media-Telecommunications/gx-deloitte-tmt-global-mobile-consumer-trends-report-first-edition.pdf
5
2.2 Drones
This section describes the use of drones, which are often referred to as vehicles that do not have
a human occupant, and when drones operate without human intervention, they are sometimes
classified as robots.
2.2.1 Air
Unmanned air vehicles (UAVs) were initially developed for military use. They have since made their
way into other uses, such as aerial photography and package delivery. UAVs are attractive, since they
can fly places manned aircraft cannot. They can also fly at low altitudes, overcoming lack of visibility
when there is cloud cover, and thus images from drones are higher resolution than satellite. The first
documented use of drones was after Hurricane Katrina in the United States of America in 2005 (Meier,
2015). Because roads were blocked by trees, small drones were deployed to search for survivors and
assess river levels.
Examples of UAV use for different disaster phases include:
(a) Preparedness such as filming volcanic activity in order to determine when warnings should be
issued (Husain, 2018).
(b) Response such as delivering equipment to locations where networks have been affected by a
disaster: For example, China has been using drones to deliver mobile gear to affected areas as
well as a virtual tower functioning as a base station (China Telecommunications Corporation,
2018). Drones are already used to deliver blood in several countries, and this could be expanded
to include other medical supplies and equipment needed during a disaster (Smyth, 2017).
Another example is the use of drones to assist Australian firefighters at night. A Lockheed Martin
Indago drone (Figure 2.2) streamed live video to operators on the ground, who used the images
to determine fire location and intensity, and find people and property that were at risk. The
drone helped save an estimated 100 homes, worth more than USD 50 million.6
(c) Assisting with recovery efforts by photographing disaster areas for damage assessments (section
3.1).
(d) The Pacific Drone Imagery Dashboard uses data from satellites and now drones for creating
maps for disaster preparedness, response and recovery.7
In order to facilitate the use of drones for crisis, the United Nations International Children’s Emergency
Fund (UNICEF) collaborated with the Government of Malawi to create a drone corridor, allowing the
private sector, academia and others to experiment with UAVs.
6
See Firefighting Case Study, available from www.lockheedmartin.com/en-us/products/procerus-technologies/indago
-unmanned-aerial-system/indago-firefighting-case-study.html (accessed 3 February 2019).
7
See UAV Imagery For Disaster Response: Pacific Drone Imagery Dashboard (PACDID), available from www.hotosm.org/
projects/uav-imagery-for-disaster-response-pacific-drone-imagery-dashboard-pacdid/(accessed 3 February 2019).
6
Figure 2.2: Indago UAV
Source: Lockheed Martin, available from www.lockheedmartin.com/en-us/products/procerus-technologies/indago-unmanned

-aerial-system.html (accessed 4 February 2019).
2.2.2 Underwater
Unmanned underwater vehicles (UUV) measure storm intensity and direction. One key difference
between UUVs and airborne drones is that GPS does not work underwater, so UUVs are tethered,
limiting their range (Meier, 2018b).
Six-foot long underwater drones are an example of UUV use (Figure 2.3) 8 that carry sensors to
measure ocean heat, salinity and density (Niiler, 2018). They were used during Hurricane Florence
in the United States of America in 2018. Sensors measured the ocean heat fuelling the hurricane,
transmitting the data to the National Weather Service. The data fill in gaps left by satellite images,
thus improving hurricane modelling. The data also enhance forecasting the intensity and route of
the hurricane, and the sensors measure salinity levels, to determine how much water from rain and
rivers is mixing into the ocean.
8
Recent figures are not available, but in 2013 the glider cost USD 125 000–150 000, depending on the configuration. In
comparison, a research vessel cost USD 35 000–100 000 per day. See Herkewitz (2013).
7
Figure 2.3: Slocum glider UUV
Source: Teledyne Marine, available from www.teledynemarine.com/slocum-glider (accessed 4 February 2019).
2.3 Big Data

Growing digitalization is creating an avalanche of data generated by sensors, closed circuit television,
mobile phones, financial transactions and Internet activities, to name just a few. While the huge
amount of data being generated is being mined by businesses for commercial purposes, Big Data
analytics also hold enormous potential for disaster management. Examples include the analysis of
social media communications during a disaster to understand the types of information and content
creators, in order to have more impact and reduce false information. Another example is the use of
financial transactions to monitor economic activity during and after a disaster, in order to improve
targeting of support efforts (section 3.3). Cellphone data have been used to monitor the movement
of the population during flooding (Telefónica, 2016). Big Data analytics are also used for analysing
information generated by sensors in IoT implementations as well as data from drones and robots.
Several disaster management-related projects have been deployed as part of the United Nations Big
Data for Development initiative (Box 2.2).
Big Data and crisis analytics is a term used to refer to the analysis of large data sets for disasters.
Due to advances in ICT and processing of large data sets, the ability to respond to disasters is at an
inflection point (Qadir et al., 2016). Big data tools can today process large amounts of crisis-related
data (e.g. user-generated, sensors) to support more effective disaster response. The role of Big Data
and crisis analytics in different disaster phases and a detailed perspective of crisis data sources and
technologies, challenges and pitfalls, as well as directions for future research, have been assembled
in a framework consisting of five nodes relevant to data scientists working in the disaster area (Figure
2.4). Data sources identifies the various generators of information, such as online activity, sensors and
so-called data exhaust (data triggered by users, through the use of cookies). Enabling technologies
refers to software and hardware that enable capture, and process and analyse the data, including AI
and machine learning. Pitfalls are the dangers of shortcomings in the data, such as false information
or non-representativeness. There are numerous challenges posed by the data, ranging from managing
the sheer volume to the rapid speed at which they are transmitted. Finally, future directions indicate
where more research is needed, such as analysis of data in real time and how to protect the crisis
analytics systems when deployed in an actual disaster.
8
Figure 2.4: Taxonomy of Big Data and crisis analytics
Source: Qadir et al., 2016.
One of the challenges of using Big Data is where to get it from. Mobile call data records can be a
rich resource for tracking population movements during a crisis (The Economist, 2014). But call data
records are held by mobile operators, who may not always be willing to share the data. This applies
equally to data generated by social media, search engines and other platforms where the owner of
the app considers it their data, and often have no formal sharing arrangement or only provide access
to a limited subset. The mobile phone industry association GSMA is working to make some of the data
collected by its members available through an IoT Big Data Ecosystem.9 The United Nations Office for
the Coordination of Humanitarian Affairs (OCHA) launched the Humanitarian Data Exchange, which
has a portal featuring around 8 000 standardized data sets from over 1 000 sources, including several
related to disasters.10
9
See IoT Big Data, available from www.gsma.com/iot/iot-big-data/(accessed 4 February 2019).
10
Available from https://data.humdata.org (accessed 4 February 2019).
9
Box 2.2: United Nations Global Pulse: Big Data and Disaster Management initiatives and tools
Global Pulse is an initiative of the United Nations Secretary-General to investigate the role
of Big Data for development. It has carried out several investigations related to disaster
management and developed various initiatives.
Crowdsourcing taxonomies to assist disaster management efforts: Communication is a

critical element in disaster management. A way to better listen to – and understand – what
people say in social media is to use data analytics to extract information relevant to priority
topics, the first step of which is to create taxonomies, or sets of keywords. However, building
a detailed taxonomy in different languages – including local dialects, jargon and alphabets
– can be labour-intensive and time-consuming. Language often represents a moving target,
with trends and changes quickly altering the nature of relevant keywords and phrases. When
such taxonomies are being used to seek information on time-critical issues such as disaster
management, it is essential that they are constantly and accurately updated.
In an attempt to address this challenge, Pulse Lab Jakarta launched Translator Gator in
2016, a language game the Lab developed to create text mining dictionaries for recognizing
sustainable development-related conversations in Indonesia. In 2017, the Lab released
Translator Gator 2 to test whether crowdsourcing can be used to inform disaster management
efforts. The aim of the project is to engage the ’wisdom of the crowd’ to create taxonomies
for disaster management for ten Association of South-East Asian Nations (ASEAN) countries
and for Sri Lanka. The expected outcomes of the project are to (a) translate disaster keywords
in multiple languages that can ultimately be used for computational research initiatives; (b)
use social media to understand the behaviours of affected population before/during/after
a disaster, and to improve communication with these communities; and (c) raise disaster-
preparedness awareness in multiple countries.
Haze Gazer: A crisis analysis tool: Forest and peatland fires, which occur on an annual basis
in Indonesia, affect the entire South-east Asia region, resulting in extensive environmental
destruction and threatening livelihoods. To better support affected populations, the
Government of Indonesia is looking for more timely and effective means of tracking and
managing the impact of fire and haze events. In response, Pulse Lab Jakarta developed Haze
Gazer, a crisis analysis and visualization tool that provides real-time situational information
from various data sources to enhance disaster management efforts. The prototype enhances
disaster management efforts by providing real-time insights on (a) the locations of fire and
haze hotspots; (b) the strength of haze in population centres; (c) the locations of the most
vulnerable cohorts of the population; and, most importantly, (d) the response strategies of
affected populations, including movement patterns and in-situ behavioural changes.
Haze Gazer uses advanced data analytics and data science to mine open data, such as fire
hotspot information from satellites and baseline information on population density and
distribution, as well as citizen-generated data, including the national complaint system in
Indonesia called LAPOR!, citizen journalism video uploads to an online news channel, and
real-time Big Data such as text-, image- and video-oriented social media.
The tool is currently being tested and improved based on feedback from disaster management
practitioners. Haze Gazer has the potential to enable Indonesia’s local (BPBD) and national
(BNPB) disaster management authorities to target their interventions and to align their
efforts with those of affected populations to increase community resilience.
10
Monitoring social response before and after disasters with data analytics: Building on
the Haze Gazer platform, CycloMon is an analytics and visualization platform developed to
assist governments in supporting communities to prepare for and respond to the impact
of tropical cyclones. The platform was developed to monitor social response before, during
and after cyclones across 14 countries in the Pacific region. CycloMon’s basic function is to
collect, analyse and visualize information from weather satellites on the path of a cyclone,
as well as insights from social media on the preparations and impact of the cyclone on
communities. The platform contains three modes to help cyclone monitoring: (a) Normal
Mode: to monitor citizens’ preparedness to cyclones across 14 countries in the Pacific region;
(b) Emergency Mode: to present social media signals from a country affected by a cyclone,
along with information on the cyclone itself; and (c) Country-Specific Mode: to provide
more detailed historical insights on disaster preparedness and impact from text-, image-
and video-based feeds.
Using mobile phone activity for disaster management during floods: In this study, mobile
phone activity data were combined with remote sensing data to understand how people
communicated during severe flooding in the Mexican State of Tabasco in 2009, in order to
explore ways that mobile data could be used to improve disaster response. The results of
the study showed that the patterns of mobile phone activity in affected locations during
and after the floods could be used as indicators of (a) flooding impact on infrastructure and
population, and (b) public awareness of the disaster. These early results demonstrated the
value of a public–private partnership in using mobile data to accurately indicate flooding
impacts in Tabasco, thus improving early warning and crisis management.
Source: United Nations Global Pulse (https://w

ww.unglobalpulse.org/projects).
2.4 Internet of Things

The use of sensors for monitoring the conditions that could trigger disasters is not new. Developments
in cloud computing, broadband wireless networks, the sensors themselves and data analysis have led
to the emergence of powerful, integrated and real-time systems referred to as the Internet of Things
(IoT). Disaster management is an ideal case for IoT applications, since sensors can send alerts about a
number of potentially dangerous situations. Tree sensors can detect if a fire has broken out by testing
temperature, moisture and carbon dioxide levels. Ground sensors can detect earth movements that
might signal earthquakes. River levels can be monitored by sensors for possible flooding (section 3.5).
One approach to IoT is the integration of sensor data with a range of other information for a multi-fac-
eted understanding of and response to disasters. After serious landslides in April 2010 that killed more
than 50 people in Rio de Janeiro, Brazil, and left thousands homeless, a City Hall Operations Centre
was built in collaboration with IBM (Centre for Public Impact, 2016). The centre operates non-stop
monitoring of various data streams generated in the city, such as security cameras, rain gauges, traffic
signal data, the electricity grid, traffic controls, GPS-equipped public transit vehicles and social media
feeds. IBM weather forecasting software uses the data and can predict emergencies up to two days
in advance.11 Social media, radio and television broadcasts and text messaging are used to inform the
public when an emergency occurs, and sirens are used in high-risk areas, alerting people to evacuate.
11
IOT In Disaster Management: Saving Lives With Early Warning, available from https://10innovations.alumniportal.com/
internet-of-things/iot-in-disaster-management-saving-lives-with-early-warning.html (accessed 4 February 2019).
11
2.5 Artificial intelligence

Software algorithms are increasingly generating valuable insights about a variety of phenomena. This
allows computers to imitate human intelligence, hence the term artificial intelligence (AI). Examples
of AI are already operational, such as voice and facial recognition, and commercialized by products
such as the IBM Watson computer system, which integrates AI into the analysis of Big Data (Box 2.3).12
Watson has been applied to disaster scenarios by having it analyse weather and census data to help
organizations prepare for a crisis and optimally allocate resources (IBM, 2012).
AI could have tremendous impact for disaster management, from potentially predicting earthquakes
to quickening recovery and response times. Humanitarian groups are hoping to speed up map creation
by using machine learning in computer software to extract objects such as buildings and roads from
aerial images.13 Considerable research effort is currently being devoted to the use of AI for detecting
and maybe one day predicting earthquakes (see section 3.6). AI does not need to be costly, as shown
in research by the Tanzania Meteorological Agency on weather and climate monitoring (Kikwasi, 2016).
The Agency used the PHP programming language to execute equations regarding meteorological
observations, with the software refining its calculations to make better predictions. The cloud-based
system features a user-friendly web-based interface, and utilizes the free open source MySQL database
management software.
Box 2.3: Artificial intelligence and machine learning in emergency situations

Artificial intelligence (AI) and machine learning have advanced to the state where they are
highly proficient in making predictions and in identification and classification.
1. Processing information: AI is used for image recognition of satellite photos to identify

damaged buildings, flooding, impassable roads, etc. Multiple data streams can be
combined with unreliable data removed and heat maps generated. For example,
DigitalGlobe (https:// w
ww.digitalglobe.com) provides open source software for disaster
response that learns how to recognize buildings on satellite photos. Following the
Nepal earthquakes in 2015, humanitarian and relief groups used pre- and post-disaster
imagery and utilized crowdsourced data analysis and machine learning to identify
locations affected by the quakes that had not yet been assessed or received aid.
2. Emergency calls: During a crisis, call centres are often overwhelmed. In addition to
voice calls, emergencies are increasingly reported by text messages and social media.
AI and machine learning are being applied to cope with the volume and different types
of calls. In the United States of America, Watson, developed by IBM, is being used for
speech-to-text recognition at emergency call centres (IBM, 2017). The text is input to
analytical software that guides operators on how to respond to the call.
3. Social media analysis: Real-time information from Facebook, Twitter, Instagram and
YouTube can be analysed and validated by AI to filter and classify information and make
predictive analysis. Artificial Intelligence for Disaster Response (AIDR) was created to
process the large number of tweets generated during a crisis.1 AIDR uses machine
learning to automatically process tweets in real time. The software collects tweets
based on hashtags and keywords, and then uses AI to further classify them by topic.
The open software is free for those who work in crisis response.
1
Available from http://aidr.qcri.org (accessed 4 February 2019).
12
Available from www.ibm.com/watson/about/index.html (accessed 4 February 2019).
13
See Integrating Machine Learning into the Tasking Manager: Notes on a Direction, available from www.hotosm.org/
updates/integrating-machine-learning-into-the-tasking-manager (accessed 4 February 2019).
12
4. Predictive analytics: AI is being used to analyse past data to predict what is likely to
happen in the event of a disaster. Optima Predict software processes information from
emergency response systems to optimize ambulance routes (Young, 2017). The data
can be integrated with online dashboards so that emergency personnel can respond
in real time.
Source: Adapted from Sugandha Lahoti (2018).
2.6 Robots
Although industrial robots have been around for some time, robots have become more sophisticated
through integration with microprocessors and sensors. The growing dexterity of robots makes them
suitable in disaster situations that are too dangerous for humans or rescue animals. Search-and-rescue
robots were reportedly first used following the September 2011 terrorist attack in New York City to
assess the wreckage of the demolished World Trade Center (Feuilherade, 2017). Since then, more
than 50 deployments of robots for disaster use have been reported.
Breakthroughs are being achieved in Japan, where there is the possibility for commercialization of
robots designed specifically for disasters (section 3.4).
2.7 Blockchain
One challenge during a rapidly evolving disaster is coordinating and verifying information among
different stakeholders. For example, the United Nations found that, in the wake of the 2010 Haitian
earthquake, assistance efforts were hampered by too many data sources among the some 20 relief
groups (Rohr, 2017). The Blockchain distributed ledger system and chain of verified records could
play a significant role in ameliorating information control.
In the United States of America Centers for Disease Control and Prevention is planning a pilot test of
blockchain for a more rapid and reliable collection of data during a crisis, in order to reduce the spread
of disease (Orcutt, 2017). This has relevance for disaster management, since similar to public health,
agencies offering relief (e.g. government, assistance agencies, telecom operators, food suppliers,
transporters, health workers and the public) need to share trusted data quickly to collaborate effec-
tively (IBM Academy of Technology, 2018).
Another way blockchain technology is already indirectly used for disaster relief is for fundraising
activities that accept cryptocurrencies (Harmes, 2018). Several organizations – including Direct Relief,
Humanity Road and Save the Children – currently accept cryptocurrencies such as Bitcoin in their
fundraising activities.
2.8 Social media
2.8.1 Twitter
Twitter is a messaging service where users can post 140 character14 messages or tweets. Users must
register for the service and, in the case of smartphones, download an app. Unlike text messages,
registered users can follow other users’ tweets. So-called ’cards‘ can be included with links to photos,
videos and other media that get around the character limit. Another feature of Twitter is that hashtags
14
As of November 2017, Twitter has increased the character count of Tweets in certain languages to make it easier to
share what is happening.
13
can be added using # to quickly receive or find tweets on a specific topic. For example, the keyword
#hurricane locates all tweets containing the keyword hurricane. The drawback is that multiple hashtags
can be created by different organizations during a disaster, making it difficult to know which one to use.
Twitter Lite is a stripped-down version of the application that is useful for many in developing nations
who only have access to slower-speed mobile networks (e.g. 2G/3G), which uses less data, thus saving
prepaid bundle usage. It is available for Android smartphones in more than 45 countries (Shah, 2018).
Usage of Twitter is not as high as other communication technologies, and recent growth has been
somewhat stagnant. There were 335 million users around the world in June 2018 (Figure 2.5), some
4.4 per cent of the world’s population.15 This limits its suitability for reaching a large number of
users during a disaster, particularly given the low penetration of Twitter outside large urban areas in
developing nations.
Figure 2.5: Global Twitter users
Source: Twitter.
Twitter has implemented the Alerts feature, aimed at use of the service during disasters.16 Alerts are
sent instantly as either tweets or converted to regular text messages. The Alerts service was launched
following the experience of the Lifeline project in Japan,17 and major public safety organizations in
Japan are participating, including the Tokyo Metropolitan Government and Fire Department, and the
Osaka Police. Organizations in other countries include the City of Seoul in the Republic of Korea; the
Rio de Janeiro Operations Centre in Brazil; and the City of Sydney, the Red Cross, the Department of
Health, and fire and police departments from several municipalities in Australia.
Studies (Baek, 2016; Kongthon et al., 2012; Takahashi et al., 2015) have looked at the role of Twitter
during disasters, primarily flooding, hurricanes and typhoons. Most of these studies deploy Big Data
analysis to examine the type of tweets and what they are used for, rather than their effectiveness
during the disaster. Therefore, the goal is primarily to investigate how Twitter is used to improve
planning and achieve more effective disaster communications. This includes understanding who are
the most influential tweeters (i.e. biggest number of followers or retweets) or categorization of tweets.
15
See Selected Company Metrics and Financials, available from https://investor.twitterinc.com/static-files/b66a0b83
-5dad-4a45-a89f-d7d4034c5287 (accessed 4 February 2019).
16
See How to use Twitter Alerts, available from https://help.twitter.com/en/managing-your-account/how-to-use-twitter
-alerts (accessed 4 February 2019).
17
See Twitter Alerts: New countries and features, available from https://blog.twitter.com/official/en_us/a/2013/twitter
-alerts-new-countries-and-features.html (accessed 4 February 2019).
14
2.8.2 Facebook
Like Twitter, Facebook also has features specific to crises. Its Crisis Response app allows users to
mark themselves as safe (Safety Check), reassuring friends and family; provide or seek help; donate
money; and receive information.18 Additionally, other features of the application – such as Groups,
Pages, Events and Fundraisers – can be useful during a crisis.19
Figure 2.6: Facebook Safety Check
Source: https://www.facebook.com/about/crisisresponse/
Facebook is relevant for disaster relief, given that it is the largest social media platform in the world,
with 2.2 billion users in June 2018 (Figure 2.7), equivalent to 30 per cent of the world’s population.
18
See www.facebook.com/about/crisisresponse/, (accessed 4 February 2019).
19
See www.facebook.com/disaster/posts/after-hurricane-florence-made-landfall-yesterday-morning-the-facebook
-community-/10156736515514936/(accessed 4 February 2019).
15
Figure 2.7: Number and distribution of Facebook users
Source: Facebook.
Facebook has a Disaster Maps tool showing where users are located, moving to and whether they
are using the Safety Check feature. The Disaster Maps tool has been used by relief organizations to
identify where Internet connectivity required restoration in Puerto Rico following Hurricane Maria, and
where respiratory masks were needed during the Southern California Wildfires (Maas et al., 2018).
The maps are particularly useful for examining mobile cellular network coverage, cellphone battery
charging and population movement during a disaster. For example, the maps were used to show 2G
network coverage after a volcanic eruption in Alotenango, Guatemala (Figure 2.8).
Figure 2.8: 2G network coverage following volcanic eruption in Guatemala
Note: The map depicts big drops in cellphone use (red and orange areas) and complete blank spots reflecting a sharp drop in
network use in the areas surrounding the site of the eruption.
Source: Facebook (https://research.fb.com/new-data-tools-for-relief-organizations-network-coverage-power-and-displacement/).
One challenge is that the maps are reflective of people who use the app and have location services
turned on. People who do not use Facebook are not included. Facebook will work with UNICEF, the
16
World Food Programme and the Red Cross to find solutions for adjusting the data to be more repre-
sentative.20
2.9 Crowdsourcing and crowdfunding

The public is increasingly using digital technologies to support disaster management. This includes
crowdsourcing (e.g. adding content to disaster maps) and crowdfunding (raising money for disaster
victims using text messaging and crowdfunding platforms).
Crowdsourcing is used to add vital details to maps of disaster areas. This enhances satellite imagery
by providing greater granularity, making relief efforts more effective and targeted. It is also quicker
and cheaper, especially since the crowd often does this on a volunteer basis compared with traditional
methods (Becker and Bendett, 2015).
Missing Maps is a project launched in 2014 and spearheaded by the American Red Cross, the British
Red Cross, Médecins Sans Frontières–UK, and the Humanitarian OpenStreetMap Team (American Red
Cross, 2015). It uses volunteers to add information to maps in order to better serve those affected by
a crisis and get assistance to them more rapidly. In the first stage, volunteers edit satellite imagery to
add details (Figure 2.9). More than 2 500 people have provided assistance, making 3.7 million edits
and locating 4.5 million people on the map. The next stage involves working with local communi-
ties to add details about roads and buildings. The last stage includes data validation and inputting
project-specific information to generate maps for different disaster programmes.
Figure 2.9: Putting communities on a map
Source: American Red Cross (2015).
Digital technologies provide another way of raising donations for disaster relief victims. More than
USD 40 million was contributed via text messaging following the Haiti earthquake disaster in 2010
(Smith, 2012). In India, Twitter was used to alert people to crowdfunding sites for those affected by
floods (section 3.2). Some relief agencies have begun accepting donations in cryptocurrencies such
20
See Facebook Disaster Maps, available from https://research.fb.com/facebook-disaster-maps-methodology/(accessed
4 February 2019).
17
as Bitcoin. Mobile money has been increasing financial inclusion in many countries and is a useful
way for friends, family and relief organizations to transfer funds to those affected by disasters (CSR
Asia, 2014). Following the 2010 Haiti earthquake, the Bill and Melinda Gates Foundation and the
United States Agency for International Development (USAID) offered a reward to the first company
to launch mobile money in the country, in order to reduce security issues involved with distributing
cash (Sossouvi, 2012).
Box 2.4: Google Crisis Map

Google has created a Crisis Map (google.org/crisismap) for users to help locate critical
emergency information. The maps feature satellite imagery and relevant information such as
the weather, flood zones, evacuation routes, shelters and power outages. Users can zoom in
on specific events such as wildfires, as shown in the map below. Crisis Map mainly shows the
situation in the United States of America, drawing on data from the United States National
Hurricane Center and weather.com. Users can request to add layers.
Source: Google (https://google.org/crisismap/us-wildfires).
2.10 Others
Traditional technologies, though not considered disruptive, continue to play a critical role in disaster
management, and are also benefitting from digitization. Radio and television broadcasting, which
now exists in digital versions, remains vital for informing the public about preparedness and response
procedures. However, they must operate in the event of a disaster, which can be an issue, especially for
television, since it generally requires a reliable electrical source.21 In many least developed countries
(LDCs), the spread of mobile communications has resulted in cellphones being more prevalent in
households than radios or televisions (Table 2.1). Therefore, cellphones may be more relevant for
communicating with the public during a crisis, provided they can be charged.
21
Radio is more resilient and digital shortwave is being promoted as a good communication solution during disasters
(Digital Radio Mondiale, 2014).
18
Table 2.1: Radio, television and mobile phone (% of households), selected LDCs
Total national (urban and rural) Rural

Country Year Radio Television Mobile Radio Television Mobile
Bangladesh 2014 3.5 43.5 88.5 3.6 33.0 86.7
Haiti 2016-17 48.0 30.7 75.9 37.3 11.5 66.5
Lao P.D.R. 2017 … 79.3 90.1 … 72.2 86.6
Madagascar 2016 48.9 17.1 34.0 46.6 12.1 29.1
Timor-Leste 2016 24.5 40.2 84.3 21.6 27.8 80.5
Yemen 2013 39.9 66.8 80.0 39.4 54.5 74.0
Source: Demographic and Health Surveys (https://dhsprogram.com).
Satellite-based emergency mapping (SEM) has improved both in quality and response (Voigt et al.,
2016). New radar apertures can see through storm clouds, and coordination between the satellite
and relief communities has improved tremendously. Satellite activations for disasters have risen from
7 in 2000 to 123 in 2014 (Figure 2.10). Challenges remain with the time required to convert satellite
images to maps and to reprogram satellite positions.
Figure 2.10: Number of disasters with satellite response
Note: EM-DAT refers to the global database on natural and technological disasters maintained by the Centre for Research on the
Epidemiology of Disasters (CRED).
Source: Voigt et al. 2016.
Radar radio waves continue to play a critical role for disaster monitoring. One example is high-fre-
quency ocean radar systems used for monitoring tsunami activity. The radar monitors the ocean
surface in real time, measuring its velocity, and computer software suggests the probability of a
tsunami (International Oceanographic Commission/United Nations Educational, Scientific and Cultural
Organization, 2018).
19
3 Case studies
This section highlights the use of disruptive technologies for disaster management in several countries.
3.1 Drones for disaster damage assessment in Vanuatu

Cyclone Pam struck the South Pacific archipelago nation of Vanuatu on 13–14 March 2015. It was
one of the worst disasters to hit Vanuatu, destroying thousands of buildings and leaving around
75 000 people homeless. Relief efforts were quickly organized, with a strong need for assessment of
the areas affected. Drones were felt to be an ideal solution for rapid and granular evaluation of the
situation, particularly since cloud cover obscured satellite images. The World Bank contracted two
drone teams from Australia and New Zealand22 as the first project of using UAVs for disaster resilience
(Bonte-Grapentin et al., 2017). The drone activities were supervised by the Humanitarian UAV Network
(UAViators).23 The Civil Aviation Authority of Vanuatu reviewed the drone use for compliance with
regulations and provided usage guidelines (Niel, 2017).
Some 200 flights were flown and the imagery taken from the drones allowed relief workers to
determine which houses were unrepairable compared with those that could be fixed, helping to guide
funding and recovery efforts. Drones are an improvement from airplane or satellite imagery, which
cannot provide as much detailed information. Because drones can collect imagery at different angles,
walls and other parts of buildings can be assessed. Agriculture is evaluated by measuring how many
crops have been destroyed and consequently how much food people will need from other sources.
Wells are examined to see if they are still working.
The imagery collected is owned by the Government of Vanuatu and published on Mapbox, an open
source mapping platform (Figure 3.1).24 This is shared with the Humanitarian OpenStreetMap Team
(Leson, 2015) and Micromappers25 to draw roads and buildings and evaluate damage to individual
structures. Volunteers uploaded and geo-tagged images from social media to overlay on the map
(Irin, 2015).
The higher resolution of damages available with drone photography provides a more accurate figure of
the cost of rebuilding compared with traditional methods. The drones also identify the most seriously
affected communities so that assistance can be prioritized.
There were challenges surrounding connectivity, the weather conditions and data formats.
Nevertheless, the experience provided valuable insights, with the drones providing the fastest method
of mapping the damaged areas (Meier and Soesilo, 2015).
22
The team from Australia used two Indago quadcopters (https://www.lockheedmartin.com/en-us/products/procerus
-technologies/indago-unmanned-aerial-system.html), while the team from New Zealand used three Alliance
hexacopters.
23
Available from http://uaviators.org (accessed 4 February 2019).
24
Available from www.mapbox.com (accessed 4 February 2019).
25
See MicroMaps of Cyclone Pam – Vanuatu – April 2015, available from https://micromappers.wordpress.com/maps/
(accessed 4 February 2019).
20
Figure 3.1: Mapbox of Cyclone Pam damage
Note: Red structures are heavily damaged, blue structures moderately damaged, and yellow structures undamaged.
Source: Matt Irwin, Mapping Cyclone Pam’s destruction with drones.
3.2 Twitter during the Chennai flood, India

Monsoons regularly affect Southern India during the months of October to December. However, in
2015 the amount of rainfall was 90 per cent above normal, attributed to the El Niño effect. It was
the worst rainfall in 100 years, causing severe flooding, which was magnified as a result of sprawling
urban development and inadequate drainage systems. It was estimated that more than 500 people
died and 1.8 million people were displaced during the several weeks of flooding (Nair et al., 2017).
Although emergency telephone numbers were established, they were overloaded, making it difficult
to get through, making social media such as Facebook and Twitter a popular method for communi-
cation for those with Internet access.
Twitter was deployed by a number of different groups for a variety of different purposes. The Twitter
live, real-time public platform was used by NGOs, the public, government agencies and the media
to share and exchange information. This resulted in unprecedented collaboration on the platform in
responding to the crisis.
Twitter India publicized three hashtags to be used during the flood, depending on the nature of
the tweet: #ChennaiRainsHelp, #ChennaiRescue and #ChennaiVolunteer for citizen groups to help
agencies on the ground (Figure 3.2). A variety of information was shared, including helpline phone
numbers, updated train schedules, weather forecasts, relief efforts and safety tips.26 This helped to
magnify critical messages, organize relief efforts, assist government agencies, warn the public and
provide information in real time to citizens trapped in the floods.
26
For examples of the Tweets, see First Post (2015).
21
Figure 3.2: Twitter India hashtags for the Chennai floods
Twitter also triggered use of other applications to assist with the crisis. For example, after a suggestion
from another Twitter user, a Google spreadsheet was created that crowdsourced information about
shelter, weather and other useful tips used by the public, police and the government. In turn, through
suggestions on Twitter, the spreadsheet was turned into a website (chennairains.org) (Lal, 2017).
Twitter was also used to raise money on crowdfunding sites. For example, hashtag #rebuildingchennai
was used to alert people of funding drives on platforms such as BitGiving and Ketto (Figure 3.3).27
27
Available from https://twitter.com/hashtag/rebuildingchennai and https://chennaifloods.ketto.org (both accessed 5
February 2019).
22
Figure 3.3: Twitter link to crowdfunding for Chennai floods
Twitter India organized a workshop in July 2016 on lessons learned from the experience and how to
have greater impact in the future. Other organizations also organized meetings post-crisis to discuss
how use of Twitter could be improved, such as optimal use of hashtags, deploying media cards, dealing
with misinformation, etc. (Kaul, 2016).
Researchers studied the use of Twitter during the flood using Big Data software to categorize tweets
and analyse the most influential tweeters based on the number sent, retweets and number of followers
(Nair et al., 2017). The results of the research are aimed at understanding the role of Twitter for
planning and managing future relief measures. The research based its analysis on the #Chennaiflood
hashtag, which was not one of the tags publicized by Twitter India. This affects the completeness
of the findings, pointing to the need for standardized hashtags for social media response during a
disaster. It is also notable that Twitter was popular due to the flood occurring in mainly urban areas,
where there is a higher level of ICT penetration and social media use. It is not clear there would be
the same impact in rural zones.
23
3.3 Financial big data for Hurricane Odile, Mexico28

Hurricane Odile struck the Mexican State of Baja California Sur 15–17 September 2014. It was one of
the strongest hurricanes ever to hit the area, and caused widespread damage.
Researchers from BBVA Data and Analytics29 and United Nation Global Pulse30 used Big Data techniques
to explore the relationship between financial transactions in the state (bank card payments and
ATM cash withdrawals) before, during and after Odile, to analyse its economic impact on those
affected. This analysis can improve understanding about which groups are most affected for targeting
post-disaster assistance, how long it takes to return to normal, and how to generate estimates of the
economic impact.
The study used data generated by more than 100 000 users of banking services in the state who
carry out around 25 000 point-of-sale and ATM cash withdrawals a day. The data were anonymized
and aggregated in conformance with Mexico laws. The data set supports analysis of the amounts and
times when transactions were made; the type of store where the purchase was made; and purchaser
demographics such as gender, age and residence postal code.
The hurricane impact is noticeable from the financial transactions. The day before the hurricane,
there was a shift in the types of transactions, with spending on food and gas increasing. The day of the
hurricane, transactions dropped dramatically, partly attributed to the blackout that occurred (Figure
3.4). Recovery time was modelled as the number of days for transactions to return to 90 per cent of
their normal level. While on average this was about two weeks, there were significant differences,
from less than two days to over a month in areas most seriously affected by the storm. Following the
storm, there were more ATM cash transactions than point-of-sale due to the large number of shops
that remained closed. Lower-income groups recovered faster (in terms of ATM transactions returning
to the normal level), likely because they had less financial capacity to stockpile goods prior to the
storm and because they spent less than wealthier groups to begin with. The data also showed that
women had longer recovery times than men.
Figure 3.4: Expected versus real banking transactions in Baja California Sur, September—October
2014
Source: BBVA, available from www.bbvadata.com/odile/(accessed 5 February 2019).
The researchers point to a number of uses for the analysis, including more precise targeting of supplies
or cash transfers to areas most economically affected by the disaster. Insights from transaction data
28
This case is adapted from: Martinez, Elena Alfaro et al. 2016.
29
BBVA is a Spanish bank with a subsidiary in Mexico (Bancomer). Available from www.bbvadata.com (accessed 5
February 2019).
30
Global Pulse is an initiative of the United Nations Secretary-General on the use of Big Data for development. Available
from www.unglobalpulse.org/about-new (accessed 5 February 2019).
24
could also be incorporated into current mechanisms for estimating economic losses. The data could
also be used to enhance inventory decisions in the future to reduce depletion of essential items. It
would also be attractive to develop an ex ante model of the study so that financial response to disasters
could be analysed in real time to provide ongoing feedback loops to relief efforts.
One limiting factor of the model is that findings are dependent on the level of financial inclusion for
monitoring transactions. Although the study found that half the Mexico population had bank accounts,
a 2017 survey shows the level to be lower, at 35 per cent, barely half the world average (Table 3.1).
The percentage of people with bank accounts rises with income levels. Therefore, unless provision is
made for the unbanked (or those using mobile money), analysis of financial transactions during and
after disasters will be unrepresentative of those likely to need the most support.
Table 3.1: Account ownership by income group, 2017
Financial institution account Mobile money account

(% age 15+) (% age 15+)
Mexico 35 6
World 67 4
Low income 24 18
Lower middle income 56 5
Upper middle income 73 3
High income 94 ..
Source: World Bank Global Financial Inclusion database.
3.4 Disaster robots in Japan

Japan is a pioneering centre for the use of robots in disasters. The Human–Robot Informatics
Laboratory of Tohoku University31 is active in developing several types of robots for disaster response:
(a) Active Scope Camera is a snake-like robot covered by small vibrating filaments, allowing the
robot to crawl over obstacles, follow walls and make turns in tight spaces. Its design makes it
useful for searching through rubble to find survivors, such as in a train wreck, and it can also
detect sounds, such as people trapped in collapsed houses.
(b) Quince is a mobile robot equipped with four sets of tracked wheels, some of which can move
up and down to allow the robot to negotiate obstacles. It carries cameras, as well as infrared
and carbon dioxide sensors, for detecting the presence of survivors trapped under rubble.
(c) Kenaf can explore a damaged building, underground facility or unstable ground after a
disaster. Kenaf enters a hole of approximately 60 cm in diameter, and is operated via wireless
communication to obtain images, sounds and 3D shapes.
(d) Search-and-rescue dogs are used for finding victims in disaster sites. However, in Japan, search-
and-rescue dogs are not used for real rescue missions due to the inability to track the animal
in real time. Research is ongoing to record and visualize the dog’s activities using cameras, GPS
and inertial measurement units, which are being designed to be small and light in order to fit
on a dog pack.
Damage to the Fukushima Daiichi Nuclear Power Plant in the 2011 earthquake triggered significant
robot research, due to radiation preventing humans from carrying out direct clean-up activities.
31
See Rescue Systems, available from www.rm.is.tohoku.ac.jp/rescue+systems/(accessed 5 February 2019).
25
In 2013, the Government of Japan established the International Research Institute for Nuclear
Decommissioning, including the mandate of developing robots. The institute consists of a consortium
of public utilities and private companies – such as Hitachi, Mitsubishi and Toshiba – which has
developed around 20 robots. There is a research and development centre near the plant where robot
operators practice on giant 3D models of the reactors as well as and on life-size copies.
One of the biggest challenges has been how to determine what happened to the fuel inside the core
of the reactor. Various robots have been used to penetrate the core but without success. Finally, in
2017, a small robot dubbed ’Little Sunfish‘ (Figure 3.5) – equipped with five propellers, video cameras,
an array of sensors and designed to operate underwater under severe radiation exposure – succeeded
in locating the missing fuel inside the reactor core (Beiser, 2018).
Figure 3.5: Little sunfish robot
Source: BBC (https://www.bbc.com/news/in-pictures-40298569).
Japanese manufacturer Honda has a robotics division known for development of the ASIMO humanoid
robot.32 It is now developing a disaster response robot, E2-DR (Figure 3.6) (Ackerman, 2017). The robot
can walk, climb over obstacles, and climb ladders and stairs. E2-DR features cameras, 3D sensors and
hand grippers. There are currently some limitations, including a battery life of one and a half hours,
and a limit of 20 minutes in wet weather. If the prototype reaches fruition, it could have a major
impact for rescue operations and clearing hazardous materials.
32
See Honda Robotics, available from https://global.honda/innovation/robotics.html (accessed 4 February 2019).
26
Figure 3.6: Honda E2-DR robot
Source: Institute of Electrical and Electronics Engineers (IEEE) (Ackerman, 2017).
3.5 Internet of Things for river flooding control in Colombia

One example of an IoT deployment is the village of Salgar, Colombia (Libelium, 2017). Following the
flooding of the Liboriana River, which triggered a devastating landslide in May 2015, causing more
than 80 deaths, the Colombia Government National Unit for Disaster Risk Management (http://
portal.gestiondelriesgo.gov.co) took steps to mitigate future occurrences. It hired a Colombia-based
company to implement an early warning system using IoT technology. Five solar-powered sensors
were installed along the Liboriana and two other rivers to monitor water levels and air tempera-
ture using ultrasound. The use of solar ensures the sensors continue to function in the event of an
electricity outage. The price of each sensor is EUR 5 200 (www.the-iot-marketplace.com/smart-water
-iot-vertical-kit). Sirens were installed in the municipality and flood hazard zones. Communications
among the sensors were a challenge, due to the hilly terrain and patchy 3G coverage. Consequently,
a 900 MHz mesh network was installed, allowing the sensors to transmit data to the control centre in
the village for local authorities to monitor and decide whether to activate the early warning system.
Data are also stored in the cloud on the eagle.io remote sensor site for others to access33. The system
software automatically sends a text message to village authorities if a risk is detected. When required,
the sirens are activated and blinking lights are displayed indicating the river level. The system is
scalable, so more sensors and monitoring points can be added. The Government of Colombia spent
COP 410 million (USD 130 000) to deploy the system (Federman, 2016).
33
https://eagle.io/
27
Figure 3.7: IoT system for river monitoring in Salgar, Colombia
Source: Libelium 2017.
Colombian researchers have investigated the event to better understand the dynamics and how to
reduce risks and loss of life from extreme rainfall events in the future, given the frequency of flash
floods and torrential rainfalls, which occur regularly in the country (Velásquez et al., 2018). A model
has been constructed of the hydrological and meteorological conditions in order to develop applica-
tions to improve risk management. Reflectivity images captured from the C-Band radar Early Warning
System of Medellín, about 60 km away, were used to classify the precipitation fields and estimate the
amount of rainfall and the cause of the landslide. Future research is planned in Colombia to determine
if satellite images can provide the same granularity in areas where there is no radar.
3.6 Artificial intelligence for earthquake detection and prediction

The volume of seismic data has increased exponentially, providing fuel for solutions to reliably detect
and locate earthquakes. Scientists have based detection on continuous seismic records, searching for
repeating signals that provide information on upcoming earthquakes. Recent advances in AI provide
the potential to detect earthquakes and perhaps even predict them.
Seismometers detect vibration in the Earth, recording the waves on a graph (waveforms) (Figure
3.8). The distances the vibrations take to reach control stations were traditionally used to locate
earthquakes through triangulation. Newer techniques use computer software to analyse waveforms
combined with other events to more precisely locate the epicentre of an earthquake (Richards et
al., 2006). This approach requires an extended time-series of waveforms that are analysed using
cross-correlation and multi-event location software.
28
Figure 3.8: Seismogram plot
Source: Incorporated Research Institutions for Seismology (http://d

s.iris.edu/gsv/tsplotForID.phtml?evid=10604800&useCache=1
&epo=0&caller=map)
Analysing many years of earthquake waveforms requires considerable computing power. One way
to reduce the processing time is to select a subset of relevant waveforms (template matching). A
new technique referred to as Fingerprint And Similarity Thresholding (FAST), reduces the processing
involved with matching templates by selecting features (fingerprints) from waveforms. The drawback
with these methods is that location information is lost.
A new approach uses AI to reduce waveform analysis processing while retaining location informa-
tion. Developed by a team of researchers at Harvard University and the Massachusetts Institute of
Technology, ConvNetQuake implements AI based on convolutional neural network computer learning
(Perol et al., 2018). The software is trained to analyse large waveform data sets and distinguish
between noise and true earthquake signals. Instead of matching waveforms to historical data, the
software filters them, eliminating the need for a library of matching templates. One important feature
of ConvNetQuake is that it preserves location information.
29
The software was tested in Oklahoma in the United States of America, where earthquakes have
become more frequent due to increased fracking and injection of underground wastewater.
ConvNetQuake detected many more earthquakes of lower magnitude than traditional techniques.
Compared with other methods, ConvNetQuake is faster and retains location information with a high
degree of accuracy (Table 3.2). The main drawback is the amount of upfront time for the software
to ’learn‘ (around one and a half hours). ConvNetQuake has the potential for very rapid earthquake
detection and location, critical for earthquake early warning.
Table 3.2: Comparison of performance of earthquake detection methods
Table 3.2: Comparison of performance of earthquake detection methods describes the perfor-
mances of three detection methods, excluding the overhead runtimes (1.5 hours of offline training
for ConvNetQuake and 47 minutes of feature extraction and database generation for Fingerprint And
Similarity Thresholding (FAST)), and explains that the computational runtimes are for the analysis of
1 week of continuous waveform data. ConvNetQuake shows a precision of 94.8%, recall of 100% and
event location accuracy of 74.6%.
Source: ConvNetQuake
While there is ongoing progress in earthquake detection, there is debate about whether prediction
is possible. Scientists at Los Alamos National Laboratory in the United States of America are using
machine learning to experiment with prediction methods (Rouet-Leduc et al., 2017). They modelled
tectonic plates using steel beams. Movement of the beams generates waveforms analysed by
computer software with a self-learning algorithm. The software eventually predicted when the beams
would rupture. The technique is now being applied to real data, which is proving challenging, due to
the large amount of extraneous ambient noise.
30
Box 3.1: Using ground sensor technology to detect earthquakes

One technique used in earthquake early warning systems is a network of specialized ground
sensors similar to IoT. Seismic sensors are deployed in earthquake-prone areas. Data are
transmitted from the sensors to a central site, where the information is processed for signals
that might indicate an earthquake is likely. This method takes longer to detect earthquakes
than a single station, but is more accurate and has greater resiliency.
The California Integrated Seismic Network ShakeAlert system uses network sensors for
earthquake detection, with warning times of a few seconds to tens of seconds. One
challenge is the cost of installing more sensors to create a more accurate and faster network.
Japan has almost nationwide coverage, whereas in California there are still a lot of blind
zones (Figure 3.9). The Japan network K-NET (Kyoshin network) consists of more than 1 000
seismographs uniformly distributed every 20 km. Data are transmitted to the National
Research Institute for Earth Science and Disaster Resilience, and made available to the
public over a website (www.kyoshin.bosai.go.jp).
Figure 3.9: Earthquake sensor density in California and Japan
Source: ShakeAlert (https://www.shakealert.org/faq/#whateew).
31
4 Challenges, relevance and sustainability

Disruptive technologies have the potential to significantly transform disaster preparation, response,
recovery and mitigation. Some technologies, such as drones and IoT, are increasingly utilized in disaster
situations, while others are still being piloted. Wider application depends both on overcoming a
number of challenges and the relevance of sustainability in different contexts.
4.1 Challenges
The use of disruptive technologies in disaster settings faces a number of challenges limiting their
impact:
(a) Skills: While some of the technologies require few skills and their use can be learned quickly, a
high level of competence is required to successfully deploy others. Drones and robots require
skilled human technicians to deploy, operate and maintain. Big Data analytics require advanced
software, powerful computers and data science expertise, and involve significant testing and
modelling, and investment in research. Many of these skill sets and supporting resources are in
short supply in developing countries.
(b) Data deluge: Growing access to ICTs and the increasing application of sensors are generating
massive volumes of data. Such Big Data has immense relevance for disaster management.
However, the growing amount of data poses challenges for data management, analysis and
verification. As one expert puts it: “The overflow of information generated during disasters
can be as paralyzing to humanitarian response as the lack of information. This flash flood of
information is often referred to as Big Data, or Big Crisis Data. Making sense of Big Crisis Data is
proving to be an impossible challenge for traditional humanitarian organizations” (www.digital
-humanitarians.com)
(c) False information: While real-time information dissemination can save lives, the rapidity by
which the data flow makes it difficult to verify, and the consequences of false information
can be deadly. There are a number of levels on which information can be false, with negative
consequences for disaster preparedness and response. First, the public tends to exaggerate
under extreme stress (Qadir et al., 2016). A study on social media use relating to the Boston
Marathon bombing in 2013 analysed 8 million unique tweets, finding that only 20 per cent
relayed accurate information; the remaining tweets either consisted of fake content or rumours
(29 per cent) or general comments and opinions (51 per cent) (Meier, 2013). A second way that
information is false is through bias, particularly with exclusive reliance on Big Data. The Google
Flu Tracker overestimated the size of the 2013 influenza pandemic by half, forecasting twice
the amount of flu-related doctor visits (Butler, 2013). Some of the factors distorting Big Data
include intentional or unintentional false crowdsourced data and how well they represent the
actual population.
(d) Legal ramifications: Disruptive technologies pose a number of regulatory and legal challenges.
Drones often need to be registered and abide by civil air regulations, particularly in crowded
urban areas. Some jurisdictions ban the use of drones, due to security concerns. Around 60 relief
organizations have developed a code of conduct related to legal and other issues concerning
drone use in humanitarian work (Box 4.1). The increasing use of Big Data for crisis analysis poses
challenges for data protection and privacy, and even more so when the data are shared cross-
border, raising issues for international research collaboration.
(e) Scale: Few disruptive technologies have scaled in the sense of widespread application with
out-of-the-box open solutions and a supporting ecosystem. As a result, it is difficult and costly
to select appropriate solutions.
(f) Costs: Investment for implementing digital solutions for disaster management can be high. While
costs of hardware such as UAVs and sensors are continuously declining, the cost of operating,
integrating and analysing the information they generate are high. Similarly, while Big Data
32
itself costs little to generate, its analysis requires specialized software and hardware and data
science expertise. Few disruptive digital solutions have achieved the scale necessary to achieve
a dramatic reduction in costs.
(g) Ownership: This revolves around several areas. One is the ability for governments and
other national organizations involved in disaster relief to own relevant disruptive technology
equipment. As noted, though dropping in costs, the price of equipment such as drones remains
costly for many developing countries. While drones are often deployed by experienced teams,
the delay between them arriving on site and a country already having the equipment could
be significant. Another aspect relates to the data ownership. Data are generated by different
disruptive technologies. Policies need to be in place regarding ownership of data generated
during disasters, including appropriate data protection and privacy regulations. A third aspect of
ownership relates to governments directing their own strategy for use of disruptive technology
for disaster management or relying on others to do so. While a lack of resources may stipulate
the latter option, in the long run, sustainability dictates that governments themselves are
best placed to know how best to utilize disruptive technologies under different contextual
environments. This is becoming particularly important for coordination purposes, given the
variety of technology tools in the hands of different groups.
(h) Readiness: Countries vary in their capabilities to properly absorb digital technologies for disaster
situations. Within countries, there are differences in access to ICTs as well as applications. This
may make it problematic for some disruptive technologies to achieve wide impact. For example,
an analysis of citizens impacted by 2014 floods in Malaysia found that mobile phones and SMS
were used most often during the flood (Aisha et al., 2015). Facebook was the most popularly
used social media application, compared with Instagram or Twitter. WhatsApp was the mobile
messaging application used most often. There was also a significant age difference: younger
users were more likely to use social media, and there was a notable inverse relationship between
age and the use of social media. No analysis was done of differences in use by gender, although
it is notable that the majority of surveys used for the study were completed by female flood
victims (63 per cent of respondents).
Box 4.1: Humanitarian UAV Code of Conduct

Work on the Humanitarian UAV Code of Conduct began in 2014. It was led by the Humanitarian
UAV Network (UAViators), with more than 60 organizations contributing including relief
agencies such as OCHA, UNICEF, the Office of the United Nations High Commissioner for
Refugees, the United Nations Disaster Assessment and Coordination, the International
Organization for Migration, the World Food Programme and the American Red Cross. The
code has 15 guidelines, reproduced below.
1. Prioritize safety above all other concerns: Humanitarian benefits should clearly
outweigh risks to persons or properties.
2. Identify the most appropriate solution: Only operate UAVs when more effective means
are not available and when humanitarian purposes are clear, such as the assessment
of needs and the response thereto. UAV missions should be informed by humanitarian
professionals and experts in UAV operations with direct knowledge of the local context.
3. Respect the humanitarian principles of humanity, neutrality, impartiality and
independence: Prioritize UAV missions based on needs and vulnerabilities, make sure
actions are not, and are not perceived as being, politically or economically influenced;
do not discriminate or make distinctions on the basis of nationality, race, gender,
religious belief, class or political opinions.
33
4. Do no harm: Assess and mitigate potential unintended consequences that UAV

operations may have on affected communities and humanitarian action.
5. Operate with relevant permissions: UAV operations must be in compliance with
relevant international and domestic law, and applicable regulatory frameworks
including customs, aviation, liability and insurance, telecoms, data protection and
others. Where national laws do not exist, operators shall adhere to the International
Civil Aviation Organization Remotely Piloted Aircraft Systems Circular 328-AN/190,
with the approval of national authorities.
6. Engage with communities: Community engagement is important and obligatory.
Developing trust and engaging local communities encourage active partnership, build
local capacities and leadership, and enhance the impact of your mission. Information
should continuously be provided to communities regarding the intent and use of UAVs.
Refer to Humanitarian UAV Community Engagement Guidelines.
7. Be responsible: Contingency plans should always be in place for unintended
consequences. UAV teams must take responsibility for and resolve any issues involving
harm to people and property, including liability.
8. Coordinate to increase effectiveness: Seek out and liaise with relevant local and
international actors and authorities. UAV teams must not interfere with and always
seek to complement formal humanitarian coordination mechanisms or operations.
9. Consider environmental implications: Operating UAVs should not pose undue risk to
the natural environment and wildlife. UAV operators must take responsibility for any
negative environmental impact their mission causes
10. Be conflict sensitive: All interventions in conflict zones become part of conflict
dynamics and can result in very serious unintended consequences, including the loss
of life. Extraordinary caution must be used in deploying UAVs in conflict zones. Refer
to Humanitarian UAV Conflict Zones Guidelines.
11. Collect, use, manage and store data responsibly: Collect, store, share and discard data
ethically using a needs-based approach, applying informed consent where possible
and employing mitigation measures where it is not. The potential for information
to put individuals or communities at risk if shared or lost must be assessed, and
measures taken to mitigate that risk (e.g. limit or cease collection or sharing). Refer
to Humanitarian UAV Data Ethics Guidelines.
12. Develop effective partnerships in preparation for and in response to crises: Work with
groups that offer complementary skill sets (humanitarian action, UAV operations, local
context, data analysis, communications) during, and preferably in advance of, crises.
Refer to Humanitarian UAV Effective Partnerships Guidelines.
13. Be transparent: Share flight activities as widely as possible, ideally publicly, as
appropriate to the context. Convey lessons or issues to communities, relevant
authorities and coordinating bodies as early as possible.
14. Contribute to learning: Carry out and share any evaluations and after action reviews
to inform the betterment of UAV use for humanitarian action.
15. Be open and collaborative: Coordination is a multistakeholder process. This means
that lessons learned and best practices on the use and coordination of UAVs in
humanitarian settings must remain open and transparent, along with any related
workshops, trainings and simulations.
34
In addition to the code of conduct, four guidelines have also been developed on Data
Protection, Community Engagement, Effective Partnerships and Conflict Sensitivity. A best
practices guide has also been drafted based on experiences in the field.
Source: Humanitarian UAV Code of Conduct.
4.2 Relevance and sustainability

The use case for disruptive technologies can be examined across a spectrum of different factors, such
as location, complexity and cost. Characteristics of the technologies in the case study settings above
are reviewed in order to understand the relevance and sustainability of different technologies under
different conditions (Table 4.1).
In terms of location, any of the technologies could, theoretically, be applied in rural as well as
urban areas, and are equally relevant in both. However, rural areas face several stumbling blocks.
Communications is one, since almost all of the technologies require high-speed networks to function
at their best, and rural areas tend to lag in fast broadband. While most technologies will still be usable,
they will not operate to full capacity. In the case of Colombia, the lack of 3G coverage led to the
deployment of a proprietary communication network, while in Vanuatu, real-time drone assessment
was hampered by a lack of mobile broadband wireless coverage. In India, flooding occurred mainly in
urban areas, where there is good cellphone coverage. It is unclear whether the use of Twitter would
have been as effective in rural areas with lower levels of coverage and smartphone penetration. The
case on Big Data will be affected by the lower level of financial inclusion in rural areas if applied to
other developing nations. The use of robots still remains largely experimental, and there was little
evidence of them being applied in rural areas at any scale. The data used for the AI case came from
monitoring Earth movements in largely rural areas, although the computation took place in urban
universities.
Technologies varied in what phase of disasters they fit in. The IoT case was used for preparedness
through monitoring hydrological conditions. Like the AI example, the Big Data case was an ex post
research study. In an unfolding disaster, it would likely be useful for the recovery phase to identify
communities that required financial assistance. In the case of AI, if it proves practical for real-time
monitoring, it could enhance mitigation and possibly preparedness by having a better understanding
of the conditions under which earthquakes occur. The robot example was used during the recovery
phase to locate radioactive fluids, although a more generalized use would also suggest robot relevance
for response. The use of drones in Vanuatu involved both response by filming damage as well as
recovery by informing relief workers about possible prioritization of assistance. Twitter was primarily
used to inform the public and coordinate communications among the relief agencies during the
response and recovery phases.
The costs of the interventions varied and were affected by donations and whether wider expenses
were considered. In the drone case, the equipment and expertise were provided by assistance
agencies. Twitter is relatively low-cost, primarily incurring user data communication charges. The Big
Data example was a research project carried out by data scientists at a bank and a United Nations
agency. In a real-world setting, it is not clear whether there would have been a cost to obtain the data
and the price of hiring data scientists to analyse them. To obtain the cost of the robot, the previous
prototypes and expenses on simulating the real-world scenario must be factored in. Going forward,
it is not yet clear what crisis-specific robots would cost. The cost of the IoT intervention was covered
by the Government of Colombia, and amounted to just over USD 100 000. The AI example required
powerful computers to carry out the analysis. Considering wider and stronger impacts, then, the
costs of deploying mobile broadband to uncovered areas should be considered. Drones, Twitter and
IoT have higher functionality if mobile broadband coverage is available.
35
Complexity of the interventions is affected by the expertise required to develop and use those inter-
ventions and the degree of functionality required. The robot and AI examples required a high level
of expertise, and were based on years of previous research and development. The Big Data example
was undertaken by data scientists. The drone case involved experienced operators, though it is likely
the skills can be transferred fairly easily. Similarly, while the IoT case required technical knowledge,
the skills involved could also be transferred. Twitter was the least complex.
There was a range of stakeholders in the cases analysed. It is notable that, in two of the three cases
where hardware solutions were implemented (robots and AI), government was involved as a funding
source and the private sector as implementer. In the case of drones, the intervention was imple-
mented by a collection of assistance partners. Twitter was the only case where there was direct public
use in addition to government and relief agencies. The case of Big Data and AI were both research
projects involving the former banks and the United Nations, and for the latter academics. The variety
of stakeholders identifies the different parties that need to be involved to realize the potential of
disruptive technology: governments, international agencies, NGOs, the private sector, the public
and researchers.
All of the technologies raise legal issues with privacy and data protection common to almost all of
them, with robots possibly posing more of an ethical than legal issue. The degree of possible legal
considerations depends on the laws in each country. Drones often need to be registered and comply
with airspace regulations, and the information they observe and collect would be subject to privacy
and possibly data protection statutes in some jurisdictions. Likewise, Twitter, Big Data, AI and IoT
might be regulated in some markets to comply with privacy and data protection laws. IoT might have
to comply with national environmental laws regarding sensor locations and visual pollution.
Table 4.1: Disruptive technology classification (based on case studies)
Drones Social Media Big Data Robots AI IoT

Urban Dependent on Likely relatively high Likely high High High Yes
regulations
Rural High except may Possibly low Possibly low Few examples High Subject to
not have supporting (depending on spread (depending on 3G coverage
infrastructure for of ICT) level of financial or installing
full functionality inclusion) communications
(e.g., 3G/4G mobile system
coverage)
Disaster phase Response, recovery Mitigation, Recovery Recovery Preparedness Preparedness

preparedness
Costs Medium Low Low High High Low to medium
Complexity Medium to high Low to disseminate; High (to analyse Medium to high (to High Medium
(qualified drone high to analyse and and interpret data) manipulate robot,
operators, establishing interpret interpret data and
and interpreting data maintain the system)
streams)
Stakeholders World Bank, NGOs, Public, NGOs, Banks, UN Government, private Academia Government,
United Nations government sector private sector
Regulatory issues Air space regulation, Some countries have Data protection Ethics Ethics, data protection Data privacy
data protection and Internet filtering and privacy and protection,
privacy environmental
considerations
Source: Based on case studies in this document.
36
5 Conclusions and recommendations

Disruptive technologies are affecting disaster management, although the pace, scope and impact
vary among the technologies. While social media platforms such as Facebook and Twitter have been
applied in a number of emergency events, Big Data, robots and AI remain largely experimental. Large-
scale impacts will require more time and investment in skills and research.
Despite this, drones and IoT are increasing in application, as experience is gained and costs fall. Older
technologies such as satellite imagery and seismometers are still the most important methods for
detecting, monitoring and accessing disasters, and text messaging has the widest reach for commu-
nicating with the public.
While there is evidence that AI can accurately predict some types of disasters before they happen,
the application of disruptive technologies today has a more incremental effect. These technologies
are refining processes by spreading critical information more quickly, improving understanding of
the causes of disasters, enhancing early warning systems, assessing damage quickly, and adding to
the knowledge base of the social behaviours and economic impacts after a crisis strikes. Disruptive
technologies are improving situational awareness by providing the crisis community with a clearer
understanding of the extent of damage and where to prioritize resources.
This section summarizes findings from the examples of disruptive technology use and makes several
recommendations, and steps are identified that governments, relief agencies, the private sector and
assistance agencies can take to maximize benefits from the opportunities highlighted in the document.
Systemization and standardization are needed to improve the application of technology interven-
tions. In terms of text messaging, it is recognized that use of a few well-publicized short code numbers
is to be encouraged and that, similarly for social media, standardized hashtags should be employed
(Government of the Philippines, 2014). This will reduce confusion among the public and magnify
impacts. Open standards will help to lower costs and ensure interoperability of different technologies.
The Common Alerting Protocol aims to standardize national and international disaster information
to enhance interoperability.34 The mobile industry has a standard for machine-to-machine communi-
cations using fourth generation wireless technology (LTE-M) to support IoT.35 IEEE has been working
on IoT and issued a draft standard on an architectural framework.36 The ITU Telecommunication
Standardization Bureau also has a Study Group carrying out work on IoT.37 The standardization should
also extend to the use of Big Data, which is currently often shrouded in opaqueness. Clear and trans-
parent sharing protocols should be implemented, including application programming interfaces.
Reach of digital technologies must be factored in to disaster management strategies. In respect to

communications among stakeholders, this includes considering the purpose and audience. While
Twitter has proved useful in crisis situations, particularly among the relief community, its penetration
is relatively low among the general public. It should also be considered that some people may not
want to use proprietary platforms, for a number of reasons, and therefore relying only on one method
may not reach all intended recipients. Research such as the case of financial Big Data described above
faces a challenge, given low levels of financial inclusion in some countries and also the need to incor-
porate all types of financial accounts (e.g. both traditional bank accounts as well as mobile money).
A global repository featuring information on how digital technologies are being applied for disaster
management would raise awareness and understanding. Hundreds of applications of disruptive
technology are underway around the world, but experiences are often buried in news articles and
research reports. While this document highlights some examples, it is the tip of the iceberg. An
34
See Sahana Alerting and Messaging Broker (SAMBRO), available from http://eden.sahanafoundation.org/wiki/
Deployments/SAMBRO (accessed 6 February 2019).
35
See Long Term Evolution for Machines: LTE-M, available from www.gsma.com/iot/mobile-iot-technology-lte-m/
36
See Internet of Things, available from https://standards.ieee.org/initiatives/iot/index.html (accessed 6 February 2019).
37
Available from www.itu.int/en/ITU-T/studygroups/2017-2020/20/Pages/default.aspx (accessed 6 February 2019).
37
information base would be useful for identifying digital interventions that have worked, who the
implementers were and other material, in order to increase understanding about which technologies
are relevant for different country circumstances and types of disasters. Some disruptive technologies
portals exist,38 but these tend to be cross-sector, whereas what is suggested here would be specifically
about their application for disaster management.
Partnerships with the private sector and academia will be critical for understanding and applying
digital technologies for disaster prediction, detection, response and relief. Numerous uses of disruptive
technologies are being developed by the private sector. In addition, the private sector controls signif-
icant amounts of personal information in Big Data sets that are of immense use for the disaster
community. Similarly, considerable relevant research is being undertaken by the academic community.
Ties between the disaster community, private sector and academia need to be strengthened.
Scaling disruptive technologies for crisis is essential to have widespread impact and be reasonably
affordable. To date, many interventions are still pilots or carried out in an ad hoc, informal manner.
Processes should exist to identify relevant use cases and scale them. This can lower costs and bring
their utility to a wider audience. Given the vast potential of disruptive technologies for disaster
management under a huge variety of different circumstances, there is a need to nurture innovation.
This is particularly relevant given the wide country contexts around the world, different types of
disasters and crisis phases. Examples from the start-up world are relevant where incubators, labs,
competitions and venture capital are used to discover, mentor and scale up promising innovations.
One example is the IBM Call for Code challenge, inviting developers to create new applications to help
better prepare for disasters (IBM, 2018)). The company has pledged USD 30 million to help coders
scale their ideas, and includes support from the Linux Foundation. One app that has been developed
under this scheme uses weather data and supply chain information to alert supermarkets, pharmacies
and other stores to increase supplies. Another example is One Concern, a California start-up founded
by Stanford University alumni that uses machine learning to predict the impact of earthquakes,
flooding and forest fires, and thus improve situational awareness.39
Training is indispensable for the disaster community to understand how to properly deploy new
and emerging digital technologies in crisis settings. Manuals are needed for different technologies.
For example, in the case of social media, this would cover hashtag guidelines, usage by public and
relief organizations, coping with fake information, etc. Training materials would ideally supplement
existing documents for text messaging (GSMA, 2013) and UAVs.40 One Concern – a start-up that uses
AI to better manage earthquakes, flooding and other disasters – provides a training module featuring
simulated disaster scenarios to train emergency relief personnel how to use the software (Shueh,
2016).
Legal ramifications of technological research and interventions for disasters need to be understood.
This is fairly straightforward in respect to specific regulations, such as registration and regulation of
drones, but more nebulous regarding data protection and privacy. One relevant dilemma is whether
the lack of data protection and privacy laws inhibit or encourage technologies that make heavy use
of personal information. While the financial Big Data described above abided by Mexican laws in
anonymizing and aggregating personal data so individuals could not be identified, it also pointed
out that significant relevant information was lost in the process. On the other hand, the case of
the Ebola epidemic in West Africa is telling41, as mobile operators refused to share call data records
due to possible legal exposure, because of a lack of clarity regarding data privacy (The Economist,
2014). Further, while Big Data may eventually be forthcoming, in the event of a disaster it is needed
immediately, so it essential to have data access and sharing protocols in place beforehand (National
38
See AI Repository, available from www.itu.int/en/ITU-T/AI/Pages/ai-repository.aspx (accessed 6 February 2019).
39
Available from www.oneconcern.com (accessed 6 February 2019).
40
The Code of Conduct for UAVs was entirely volunteer-driven with no funding. Available from https://uavcode.org
41
See https://www.itu.int/en/ITU-D/Emergency-Telecommunications/Pages/Publications.aspx for reports on CDR
analysis in Guinea, Liberia, and Sierra Leone.
38
Security Telecommunications Advisory Committee, 2016). The disaster community has developed
codes of conduct in certain areas, including data protection, that can help when laws are vague, such
as blurring images of people when filmed by drones.42
Adequate capacity remains fundamental for properly planning and deploying relevant digital technol-
ogies. While digital technologies show great promise for all phases of disasters, on-the-ground
planning, management and operations are critical to their success (Teutsch, 2010). Disaster agencies
do not need to be experts in digital technologies, but they do need to understand enough about them
to develop proactive blueprints for deploying them. Disaster agencies might also consider creating a
chief technologist post to better understand how to apply disruptive technologies.
42
See Humanitarian UAV Code of Conduct, available from https://uavcode.org/further-guidance/131-2/(accessed 6
February 2019).
39
Abbreviations
5G Fifth Generation
AI Artificial Intelligence
AIDR Artificial Intelligence for Disaster Response
BDT Telecommunication Development Bureau
CRED Centre for Research on the Epidemiology of Disasters
GPS Global Positioning System
ICT Information and Communication Technology
IEEE Institute of Electrical and Electronics Engineers
IoT Internet of Things
LDC Least Developed Country
LSE Least Developed Countries, Small Island Developing States and Emergency
Telecommunications Division (ITU)
NGO Non-Governmental Organization
OCHA United Nations Office for the Coordination of Humanitarian Affairs
SMS Short Message Service
UAV Unmanned Air Vehicle
UNICEF United Nations International Children’s Emergency Fund
USD United States Dollar
UUV Unmanned Underwater Vehicle
40
References
Ackerman, E. (2017). Honda Unveils Prototype E2-DR Disaster Response Robot. IEEE Spectrum, 2
October. Available from https:// s pectrum.ieee.org/automaton/robotics/humanoids/iros-2017-honda
-unveils-prototype-e2dr-disaster-response-robot (accessed 5 February 2019).
Aisha, T.S., S. Wok, A. Maznina Manaf and R. Ismail (2015). Exploring the Use of Social Media During
the 2014 Flood in Malaysia. Procedia – Social and Behavioral Sciences, vol. 211 (25 November), pp.
931–937. 2nd Global Conference on Business and Social Sciences, 17–18 September, Bali, Indonesia.
Available from https://doi.org/10.1016/j.sbspro.2015.11.123 (accessed 5 February 2019).
American Red Cross (2015). Case Study: Lessons learned from community mapping in urban and
rural areas in East Africa and Bangladesh. Available from www.preparecenter.org/sites/default/files/
case_study_community_mapping_aug6.pdf (accessed 4 February 2019).
Baek, S. (2016). Crisis Communication in Major Disaster Using Natural Language Processing. Doctoral
dissertation. Available from https:// r epository.kulib.kyoto-u.ac.jp/dspace/bitstream/2433/215520/2/
dkogk04149.pdf (accessed 4 February 2019).
Becker, D. and S. Bendett (2015). Crowdsourcing Solutions for Disaster Response: Examples and
Lessons for the US Government. Procedia Engineering, vol. 107, pp. 27–33. Available from https://
doi.org/10.1016/j.proeng.2015.06.055 (accessed 4 February 2019).
Beiser, V. (2018). The Robot Assault on Fukushima. Wired, 26 April. Available from www.wired.com/
story/fukushima-robot-cleanup/(accessed 5 February 2019).
Bonte-Grapentin, M., P. Meier and K. Saito (2017). Lessons From Mapping Geeks: How Aerial
Technology is Helping Pacific Island Countries Recover From Natural Disasters. World Bank Blog. 20
November. Available from http://blogs.worldbank.org/eastasiapacific/lessons-mapping-geeks-how
-aerial-technology-helping-pacific-island-countries-recover-natural (accessed 4 February 2019).
Butler, D. (2013). When Google Got Flu Wrong. Nature News, vol. 494, issue 7436, p. 155. Available
from https:// doi.org/10.1038/494155a (accessed 5 February 2019).
Centre for Public Impact (2016). Rio de Janeiro’s centre of operations: COR. Available from www
.centreforpublicimpact.org/case-study/ioe-based-rio-operations-center (accessed 4 February 2019).
Centre for Research on the Epidemiology of Disasters (CRED), Institute of Health and Society and
Université catholique de Louvain (2016). Annual Disaster Statistical Review 2016: The numbers and
trends. Brussels.
China Telecommunications Corporation (2018). Operators’ consideration of disaster preparedness,

disaster reduction and disaster response. ITU-D SG 02 Contribution, 21 March. Available from www
.itu.int/md/D18-SG02-C-0056/en (accessed 3 February 2019).
Christensen, C.M., M.E. Raynor and R. McDonald (2015). What Is Disruptive Innovation? Harvard
Business Review, December. Available from https://hbr.org/2015/12/what-is-disruptive-innovation
CRED (2018). Natural disasters in 2017: Lower mortality, higher cost. Cred Crunch, March. Available
from www.emdat.be/publications (accessed 2 February 2019).
CSR Asia (2014). Electronic Cash Transfers In Disaster Response – Opportunities For Business
Engagement. CSR Asia Business Briefing, September. Available from www.csr-asia.com/report/CSR
-FA2.pdf (accessed 4 February 2019).
Digital Radio Mondiale (2014). Digital Radio Mondiale (DRM) To Participate With A Special Broadcast
In The International Radio For Disaster Relief (IRDR) Trial On June 5th and 6th. Press Release. 4 June.
41
Available from www.drm.org/digital-radio-mondiale-drm-special-broadcast-in-the-international-radio

-for-disaster-relief-irdr-trial-on-june-5th-and-6th-2/(accessed 4 February 2019).
The Economist (2014). Ebola and Big Data: Waiting on hold. 27 October.
Federman (2016). En funcionamiento nuevo sistema de alerta temprana SAT de SaLgar – Antioquia.
Available from https://federman.com.co/news/show/title/en-funcionamiento-nuevo-sistema-de
-alerta-temprana-sat-de-salgar-antioquia (accessed 5 February 2019).
Feuilherade, P. (2017). Robots to the rescue. IEC e-tech, April. Available from https://iecetech.org/
issue/2017-04/Robots-to-the-rescue (accessed 2 February 2019).
First Post (2015). Chennai rains: How Twitter users are making a difference to those affected by the
disaster. 2 December. Available from www.firstpost.com/india/chennai-rains-how-twitter-users-are
-making-a-difference-to-those-affected-by-the-disaster-2529478.html (accessed 5 February 2019).
Global Mobile Suppliers Association (2019). Global 5G Status – Snapshot: January 2019. Available from
https://gsacom.com/paper/global-5g-status-snapshot-january-2019/(accessed 3 February 2019).
Government of the Philippines (2014). Standard Hashtags for Disaster Response. Available from
https://reliefweb.int/sites/reliefweb.int/files/resources/pcdspo-report-on-disaster-response-unified
-hashtags.pdf (accessed 6 February 2019).
GSMA (2013). Towards a Code of Conduct: Guidelines for the Use of SMS in Natural Disasters. Available
from www.gsma.com/mobilefordevelopment/programme/mobile-for-humanitarian-innovation/
towards-a-code-of-conduct-guidelines-for-the-use-of-sms-in-natural-disasters (accessed 3 February
2019).
Harmes, D. (2018). International Charities Embrace Blockchain for Disaster Relief. Linkedin. 22 April.
Available from www.linkedin.com/pulse/international-charities-embrace-blockchain-disaster-harmes
-ph-d-/(accessed 4 February 2019).
Herkewitz, W. (2013). Ocean Drones Plumb New Depths. The New York Times, 11 November. Available
from www.nytimes.com/2013/11/12/science/earth/ocean-drones-plumb-new-depths.html (accessed
3 February 2019).
Husain, F. (2018). Drone Monitoring of Volcanoes Could Improve Warning Times. PBS, 5 June. Available
from www.pbs.org/wgbh/nova/article/drone-monitoring-of-volcanoes-could-improve-warning-times/
IBM (2012). First Academic Case Competition Proposes Novel Ways to Put IBM Watson to Work.
News Release. 17 May. Available from https://www-03.ibm.com/press/us/en/pressrelease/37733
.wss (accessed 4 February 2019).
(2017). Association of Public-Safety Communications Officials (APCO) International Taps IBM

Watson Capabilities to Evaluate and Report on Emergency Calls. News release. 14 August. Available
from https://www-03.ibm.com/press/us/en/pressrelease/53000.wss (accessed 4 February 2019).
(2018). IBM Leads ‘Call for Code’ to Use Cloud, Data, AI, Blockchain for Natural Disaster Relief.
News release. 24 May 24. Available from https:// n ewsroom.ibm.com/2018-05-24-IBM-Leads-Call-for
-Code-to-Use-Cloud-Data-AI-Blockchain-for-Natural-Disaster-Relief. (accessed 6 February 2019).
IBM Academy of Technology (2018). Disaster Management using Blockchain and IoT. IBM Academy of
Technology Blog. 1 April 2018. Available from www.ibm.com/blogs/academy-of-technology/disaster
-management-using-blockchain-iot (accessed 4 February 2019).
42
International Oceanographic Commission/United Nations Educational, Scientific and Cultural

Organization (2018). Advances in Tsunami Warning to Enhance Community Responses. Available
from http://unesdoc.unesco.org/images/0026/002628/262829E.pdf (accessed 4 February 2019).
Irin (2015). Drones and crisis mapping – digital aid in Vanuatu. Irin News. 14 April 2015. Available
from www.irinnews.org/fr/node/255217 (accessed 4 February 2019).
ITU (2017). Question 5/2: Utilization of telecommunications/ICTs for disaster preparedness, mitigation
and response, Final Report. Available from www.itu.int/dms_pub/itu-d/opb/stg/D-STG-SG02.05.1
-2017-PDF-E.pdf (accessed 4 February 2019).
(2018). Setting the Scene for 5G: Opportunities & Challenges. Available from www.itu.int/
en/ITU-D/Documents/ITU_5G_REPORT-2018.pdf (accessed 12 February 2019).
Kaul, M. (2016). Twitter for crisis and disaster relief. Twitter Blog. 28 July. Available from https://blog
.twitter.com/official/en_us/a/2016/twitter-for-crisis-and-disaster-relief.html (accessed 6 February
2019).
Kikwasi, W.K. (2016). Improving Efficiency and Quality in Weather Observation and Climate Monitoring
by Using Artificial Intelligence and Information Communication Technology (ICT) Infrastructure.
World Meteorological Technical Conference on Meteorological and Environmental Instruments and
Methods of Observation (CIMO TECO 2016). Available from www.wmo.int/pages/prog/www/IMOP/
publications/IOM-125_TECO_2016/Session_4/P4(20)_Kikwasi.pdf (accessed 4 February 2019).
Kongthon, A., C. Haruechaiyasak, J. Pailai and S. Kongyoung (2012). The Role of Twitter during a Natural
Disaster: Case Study of 2011 Thai Flood. 2012 Proceedings of PICMET ’12: Technology Management
for Emerging Technologies: pp. 2227–2232. Available from https:// ieeexplore.ieee.org/document/
6304238/(accessed 4 February 2019).
Lahoti, S. (2018). AI to the rescue: 5 ways machine learning can assist during emergency situations.
Packt. 16 January. Available from https://hub.packtpub.com/ai-rescue-5-ways-machine-learning-can
-assist-emergency-situations/(accessed 4 February 2019).
Lal, A. (2017). How social media showed its unique power of crowdsourcing during the Chennai floods.
Scroll.in, 27 November. Available from https://scroll.in/article/859304/how-social-media-showed-its
-unique-power-of-crowdsourcing-during-the-chennai-floods (accessed 5 February 2019).
Leson, H. (2015). Mapping Vanuatu Islands. Humanitarian OpenStreetMap Team News, 16 March.
Available from www.hotosm.org/updates/2015-03-16_mapping_vanuatu_islands (accessed 4
February 2019).
Libelium (2017). Early warning system to prevent floods and allow disaster management in Colombian
rivers. 12 December. Available from www.libelium.com/early-warning-system-to-prevent-floods-and
-allow-disaster-management-in-colombian-rivers/(accessed 5 February 2019).
Maas P., L. McGorman, C. Nayak, W. Park and A. Gros (2018). New data tools for relief organizations:
network coverage, power and displacement. Facebook Research. 6 August. Available from https://
research.fb.com/new-data-tools-for-relief-organizations-network-coverage-power-and-displacement/
Martinez, E.A., M. Hernandez Rubio, R. Maestre Martinez, J. Murillo Arias, D. Patane, A. Zerbe, R.
Kirkpatrick, M. Luengo-Oroz and A. Zerbe (2016). Measuring Economic Resilience to Natural Disasters
with Big Economic Transaction Data. Cornell University. 28 September. Available from http:// a rxiv.org/
abs/1609.09340 (accessed 5 February 2019).
Meier, P. (2013). Analyzing Fake Content on Twitter During Boston Marathon Bombings. iRevolu-
tions, 20 October. Available from https:// irevolutions.org/2013/10/20/fake-content-on-twitter-boston
-marathon/(accessed 5 February 2019).
43
(2015). Rescue Robotics: An Introduction. iRevolutions, 10 August. Available from https://

irevolutions.org/2015/08/10/rescue-robotics-introduction/(accessed 3 February 2019).
(2018a). Humanitarian UAV/Done Missions: Towards Best Practices. Available from https:/
/humanitariandronecode.files.wordpress.com/2018/04/humanitarian-uav-drone-missions-towards
-best-practices.pdf (accessed 12 February 2019).
(2018b). Testing Underwater Drones: Lessons Learned from the South Pacific. iRevolutions,
28 March. Available from https:// irevolutions.org/2018/03/28/underwater-drones-lessons (accessed
3 February 2019).
Meier, P. and R. Munro (2010). The Unprecedented Role of SMS in Disaster Response: Learning
from Haiti. SAIS Review, vol. 30, No. 2 (Summer–Fall), pp. 91–103. Available from https://courses
.cs.washington.edu/courses/cse490d/13wi/ClassMaterials/papers/meier_munro.pdf (accessed 3
February 2019).
Meier, P. and D. Soesilo (2015). Using Drones for Disaster Damage Assessments in Vanuatu. Drones in
Humanitarian Action. Available from https:// e uropa.eu/capacity4dev/innov-aid/document/case-study
-no-10-using-drones-disaster-damage-assessments-vanuatu (accessed 4 February 2019).
Nair, M.R. G.R. Ramya and P. Bagavathi Sivakumar (2017). Usage and analysis of Twitter during 2015
Chennai flood towards disaster management. Procedia Computer Science, vol. 115, pp. 350–358.
Available from https://doi.org/10.1016/j.procs.2017.09.089 (accessed 4 February 2019).
National Security Telecommunications Advisory Committee (2016). Report to the President on Big
Data Analytics. Available from www.hsdl.org/?abstract&did=792598 (accessed 6 February 2019).
Niel, J. (2017). Using Drones (UAS) For Disaster Damage Assessment in Vanuatu. Regional Meeting
on Broadband Connectivity for the Asia–Pacific Least Developed Countries. Port Vila. 26–27 October.
Available from http:// u
nohrlls.org/event/ldcs-regional-meeting-vanuatu/(accessed 4 February 2019).
Niiler, E. (2018). Underwater Drones Track Hurricane Florence’s Trajectory. Wired, 13 September.
Available from www.wired.com/story/hurricane-florence-underwater-drone-slocum-glider/(accessed
4 February 2019).
Orcutt, M. (2017). Why the CDC Thinks Blockchain Can Save Lives. MIT Technology Review, 2
October. Available from www.technologyreview.com/s/608959/why-the-cdc-wants-in-on-blockchain/
Perol, T., M. Gharbi and M. Denolle (2018). Convolutional Neural Network for Earthquake Detection
and Location. Science Advances, 4.2:e1700578. Available from www.ncbi.nlm.nih.gov/pmc/articles/
PMC5817932/(accessed 5 February 2019).
Poblet, M., E. García-Cuesta and P. Casanovas (2014). Crowdsourcing Tools for Disaster Management: A
Review of Platforms and Methods. In AI Approaches to the Complexity of Legal Systems. Available from
https:// p
dfs.semanticscholar.org/0b47/d7e3da7438a02cec53969b37963a29c328b0.pdf (accessed 3
February 2019).
Qadir, J., A. Ali, R. ur Rasool, A. Zwitter, A. Sathiaseelan and J. Crowcroft (2016). Crisis Analytics: Big
Data-Driven Crisis Response. Journal of International Humanitarian Action, vol. 1, No.1, p. 12. Available
from https://doi.org/10.1186/s41018-016-0013-9 (accessed 4 February 2019).
Richards, P., F. Waldhauser, D. Schaff and Won-Young Kim (2006). The Applicability of Modern Methods
of Earthquake Location. Pure and Applied Geophysics. Available from www.ldeo.columbia.edu/
~richards/my_papers/PAGEOPH_2006_eqloc.pdf (accessed 5 February 2019).
44
Rohr, J. (2017). Blockchain for Disaster Relief: Creating Trust Where It Matters Most. SAP News,
14 November. Available from https://news.sap.com/2017/11/blockchain-disaster-relief (accessed 4
February 2019).
Rouet‐Leduc, B., C. Hulbert, N. Lubbers, K. Barros, C.J. Humphreys and P.A. Johnson (2017). Machine
learning predicts laboratory earthquakes. Geophysical Research Letters, vol. 44, issue 18, pp.
9276–9282. Available from https://doi.org/10.1002/2017GL074677 (accessed 5 February 2019).
Shah, J. (2018). Twitter Lite in the Google Play Store: now available in 45+ countries. Twitter
Blog. 13 August. Available from https://blog.twitter.com/official/en_us/topics/company/2018/
twitterlitenewcountries.html (accessed 4 February 2019).
Shueh, J. (2016). One Concern: Applying Artificial Intelligence to Emergency Management. Government
Technology, 23 March. Available from www.govtech.com/public-safety/One-Concern-Applying
-Artificial-IntelligenceI-to-Emergency-Management.html (accessed 6 February 2019).
Smith, A. (2012). Real Time Charitable Giving. Pew Research Center. Available from www.pewinternet
.org/2012/01/12/real-time-charitable-giving/(accessed 4 February 2019).
Smyth, J.M. (2017). Drones and the Future of Disaster Response. Open Health News, 14 November.
Available from www.openhealthnews.com/articles/2017/drones-and-future-disaster-response
Sossouvi, K. (2012). Innovation in emergencies: the launch of mobile money in Haiti. Humanitarian
Exchange. Humanitarian Practice Network, May. Available from https://odihpn.org/magazine/
innovation-in-emergencies-the-launch-of-%C2%91mobile-money%C2%92-in-haiti/ (accessed 4
February 2019).
Takahashi, B., E.C. Tandoc and C. Carmichael (2015). Communicating on Twitter during a Disaster:
An Analysis of Tweets during Typhoon Haiyan in the Philippines. Computers in Human Behavior, vol.
50 (September), pp. 392–398. Available from https://doi.org/10.1016/j.chb
.2015.04.020 (accessed
4 February 2019).
Telefónica (2016). How big mobile data can help manage natural disasters. 27 June. Available from
www.telefonica.com/en/web/responsible-business/article/-/blogs/how-big-mobile-data-can-help
-manage-natural-disasters (accessed 4 February 2019).
Teutsch, K. (2010). Effective Disaster Management Strategies in the 21st Century. Emergency
Management, 4 August. Available from www. govtech.com/em/disaster/Effective-Disaster
-Management-Strategies.html (accessed 4 February 2019).
UNICEF (2017). Humanitarian drone corridor launched in Malawi. 23 August. Available from www
.unicef.org/stories/humanitarian-drone-corridor-launched-malawi (accessed 3 February 2019).
Velásquez, N., C. Hoyos, J.I. Vélez and E. Zapata (2018). Reconstructing the Salgar 2015 Flash Flood
Using Radar Retrievals and a Conceptual Modeling Framework: A Basis for a Better Flood Generating
Mechanisms Discrimination. Hydrology and Earth System Sciences Discussions, September, pp.
1–36. 10.5194/hess-2018-452. Available from www.researchgate.net/publication/327870499
_Reconstructing_the_Salgar_2015_Flash_Flood_Using_Radar_Retrievals_and_a_Conceptual
_Modeling_Framework_A_Basis_for_a_Better_Flood_Generating_Mechanisms_Discrimination
Voigt, S., F. Giulio-Tonolo, J. Lyons, J. Kučera, B. Jones, T. Schneiderhan, G. Platzeck, K. Kaku, M.

Hazarika, L. Czaran, S. Li, W. Pedersen, G. Kadiri James, C. Proy, D. Macharia Muthike, J. Bequignon
and D. Guha-Sapir (2016). Global trends in satellite-based emergency mapping. Science, vol. 353, No.
6296, pp. 247–252. Available from www.researchgate.net/publication/305324429_Global_trends_in
_satellite-based_emergency_mapping (accessed 4 February 2019).
45
Young, A. (2017). Can Predictive Analytics Help Avert Pittsburgh’s Next Disaster? Government
Technology, 28 April. Available from www.govtech.com/data/Can-Predictive-Analytics-Help-Avert
-Pittsburghs-Next-Disaster.html (accessed 4 February 2019).
46
International Telecommunication Union (ITU)
Telecommunication Development Bureau (BDT)
Office of the Director
Place des Nations
CH-1211 Geneva 20 – Switzerland
Email: bdtdirector@itu.int
Tel.: +41 22 730 5035/5435
Fax: +41 22 730 5484
Deputy to the Director and Infrastructure Enabling Innovation and Partnership Project Support and Knowledge
Director，Administration and Environmnent and Department (IP) Management Department (PKM)
Operations Coordination e-Applications Department (IEE)
Department (DDR)
Email: bdtdeputydir@itu.int Email: bdtiee@itu.int Email: bdtip@itu.int Email: bdtpkm@itu.int
Tel.: +41 22 730 5784 Tel.: +41 22 730 5421 Tel.: +41 22 730 5900 Tel.: +41 22 730 5447
Fax: +41 22 730 5484 Fax: +41 22 730 5484 Fax: +41 22 730 5484 Fax: +41 22 730 5484
Africa
Ethiopia Cameroon Senegal Zimbabwe
International Telecommunication Union internationale des Union internationale des International Telecommunication
Union (ITU) télécommunications (UIT) télécommunications (UIT) Union (ITU)
Regional Office Bureau de zone Bureau de zone Area Office
P.O. Box 60 005 Immeuble CAMPOST, 3e étage 19, Rue Parchappe x Amadou TelOne Centre for Learning
Gambia Rd., Leghar ETC Building Boulevard du 20 mai Assane Ndoye Corner Samora Machel and
3rd floor Boîte postale 11017 Immeuble Fayçal, 4e étage Hampton Road
Addis Ababa – Ethiopia Yaoundé – Cameroun B.P. 50202 Dakar RP P.O. Box BE 792 Belvedere
Dakar – Sénégal Harare – Zimbabwe
Email: itu-addis@itu.int Email: itu-yaounde@itu.int Email: itu-dakar@itu.int Email: itu-harare@itu.int

Tel.: +251 11 551 4977 Tel.: + 237 22 22 9292 Tel.: +221 33 849 7720 Tel.: +263 4 77 5939
Tel.: +251 11 551 4855 Tel.: + 237 22 22 9291 Fax: +221 33 822 8013 Tel.: +263 4 77 5941
Tel.: +251 11 551 8328 Fax: + 237 22 22 9297 Fax: +263 4 77 1257
Fax: +251 11 551 7299
Americas
Brazil Barbados Chile Honduras
União Internacional de International Telecommunication Unión Internacional de Unión Internacional de
Telecomunicações (UIT) Union (ITU) Telecomunicaciones (UIT) Telecomunicaciones (UIT)
Regional Office Area Office Oficina de Representación de Área Oficina de Representación de Área
SAUS Quadra 06, Bloco “E” United Nations House Merced 753, Piso 4 Colonia Palmira, Avenida Brasil
11º andar, Ala Sul Marine Gardens Casilla 50484, Plaza de Armas Ed. COMTELCA/UIT, 4.º piso
Ed. Luis Eduardo Magalhães (Anatel) Hastings, Christ Church Santiago de Chile – Chile P.O. Box 976
70070-940 Brasilia, DF – Brazil P.O. Box 1047 Tegucigalpa – Honduras
Bridgetown – Barbados
Email: itubrasilia@itu.int Email: itubridgetown@itu.int Email: itusantiago@itu.int Email: itutegucigalpa@itu.int

Tel.: +55 61 2312 2730-1 Tel.: +1 246 431 0343/4 Tel.: +56 2 632 6134/6147 Tel.: +504 22 201 074
Tel.: +55 61 2312 2733-5 Fax: +1 246 437 7403 Fax: +56 2 632 6154 Fax: +504 22 201 075
Fax: +55 61 2312 2738
Arab States Asia and the Pacific CIS countries

Egypt Thailand Indonesia Russian Federation
International Telecommunication International Telecommunication International Telecommunication International Telecommunication
Union (ITU) Union (ITU) Union (ITU) Union (ITU)
Regional Office Regional Office Area Office Regional Office
Smart Village, Building B 147, 3rd floor Thailand Post Training Center, 5th Sapta Pesona Building, 13th floor 4, Building 1
Km 28 Cairo – Alexandria Desert Road floor, JI. Merdan Merdeka Barat No. 17 Sergiy Radonezhsky Str.
Giza Governorate 111 Chaengwattana Road, Laksi Jakarta 10001 – Indonesia Moscow 105120
Cairo – Egypt Bangkok 10210 – Thailand Russian Federation
Mailing address Mailing address: Mailing address:

P.O. Box 178, Laksi Post Office c/o UNDP – P.O. Box 2338 P.O. Box 25 – Moscow 105120
Laksi, Bangkok 10210 – Thailand Jakarta 10001 – Indonesia Russian Federation
Email: itucairo@itu.int Email: itubangkok@itu.int Email: itujakarta@itu.int Email: itumoskow@itu.int

Tel.: +202 3537 1777 Tel.: +66 2 575 0055 Tel.: +62 21 381 3572 Tel.: +7 495 926 6070
Fax: +202 3537 1888 Fax: +66 2 575 3507 Tel.: +62 21 380 2322 Fax: +7 495 926 6073
Tel.: +62 21 380 2324
Fax: +62 21 389 05521
Europe
Switzerland
International Telecommunication
Union (ITU)
Telecommunication Development
Bureau (BDT)
Europe Unit (EUR)
Place des Nations
CH-1211 Geneva 20 – Switzerland
Switzerland
Email: eurregion@itu.int
Tel.: +41 22 730 5111
ITUGET 2019 Background document
Emergency Telecommunications
Disruptive technologies
and their use in disaster risk
2019
International Telecommunication Union Disruptive technologies and their use in disaster risk reduction and management 2019
Telecommunication Development Bureau
Place des Nations
CH-1211 Geneva 20
Switzerland
Photo credits: Shutterstock

Disruptive Technologies

Uploaded by

Copyright:

Available Formats

Disruptive Technologies

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Disruptive Technologies

Uploaded by

Copyright:

Available Formats

ITUGET 2019 Background document

Please consider the environment before printing this report.

An avalanche of data is being generated by sensors, closed-circuit television, smartphones, financial

2 Disruptive technologies for disaster risk reduction and management 3

4 Challenges, relevance and sustainability 32

5 Conclusions and recommendations 37

Figure 1.1: Annual disasters and economic impact

Introduction, scope, and purpose

Figure 1.2: Disaster management phases

Figure 1.3: Natural disaster subgroup

2 Disruptive technologies for disaster risk reduction and

Box 2.1: 5G and disruptive technologies

In early 2012, ITU embarked on a programme to develop IMT (International Mobile

2.1 Mobile phones

Source: Poblet et al. 2014.

Examples of UAV use for different disaster phases include:

Figure 2.2: Indago UAV

Source: Lockheed Martin, available from www​.lockheedmartin​.com/​en​-us/​products/​procerus​-technologies/​indago​-unmanned​

Figure 2.3: Slocum glider UUV

Source: Teledyne Marine, available from www​.teledynemarine​.com/​slocum​-glider (accessed 4 February 2019).

2.3 Big Data

Figure 2.4: Taxonomy of Big Data and crisis analytics

Source: Qadir et al., 2016.

Crowdsourcing taxonomies to assist disaster management efforts: Communication is a

Source: United Nations Global Pulse (https:​/​/w

2.4 Internet of Things

2.5 Artificial intelligence

Box 2.3: Artificial intelligence and machine learning in emergency situations

1. Processing information: AI is used for image recognition of satellite photos to identify

2.8 Social media

Figure 2.5: Global Twitter users

Figure 2.6: Facebook Safety Check

Figure 2.7: Number and distribution of Facebook users

Figure 2.8: 2G network coverage following volcanic eruption in Guatemala

2.9 Crowdsourcing and crowdfunding

Figure 2.9: Putting communities on a map

Source: American Red Cross (2015).

Box 2.4: Google Crisis Map

Source: Google (https:​/​/​google​.org/​crisismap/​us​-wildfires).

Total national (urban and rural) Rural

Haiti 2016-17 48.0 30.7 75.9 37.3 11.5 66.5

Lao P.D.R. 2017 … 79.3 90.1 … 72.2 86.6

Madagascar 2016 48.9 17.1 34.0 46.6 12.1 29.1

Timor-Leste 2016 24.5 40.2 84.3 21.6 27.8 80.5

Yemen 2013 39.9 66.8 80.0 39.4 54.5 74.0

Source: Demographic and Health Surveys (https:​/​/​dhsprogram​.com).

Figure 2.10: Number of disasters with satellite response

Source: Voigt et al. 2016.

3.1 Drones for disaster damage assessment in Vanuatu

Figure 3.1: Mapbox of Cyclone Pam damage

Source: Matt Irwin, Mapping Cyclone Pam’s destruction with drones.

3.2 Twitter during the Chennai flood, India

Figure 3.2: Twitter India hashtags for the Chennai floods

Figure 3.3: Twitter link to crowdfunding for Chennai floods

3.3 Financial big data for Hurricane Odile, Mexico28

Source: BBVA, available from www​.bbvadata​.com/​odile/​(accessed 5 February 2019).

Table 3.1: Account ownership by income group, 2017

Financial institution account Mobile money account

Lower middle income 56 5

2 Disruptive technologies for disaster risk reduction and management 3

4 Challenges, relevance and sustainability 32

5 Conclusions and recommendations 37

Source: Lockheed Martin, available from www.lockheedmartin.com/en-us/products/procerus-technologies/indago-unmanned

Source: Teledyne Marine, available from www.teledynemarine.com/slocum-glider (accessed 4 February 2019).

Source: United Nations Global Pulse (https://w

Source: Google (https://google.org/crisismap/us-wildfires).

Source: Demographic and Health Surveys (https://dhsprogram.com).

Source: BBVA, available from www.bbvadata.com/odile/(accessed 5 February 2019).

Source: BBC (https://www.bbc.com/news/in-pictures-40298569).

Source: Incorporated Research Institutions for Seismology (http://d

Source: ShakeAlert (https://www.shakealert.org/faq/#whateew).

Available from www.drm.org/digital-radio-mondiale-drm-special-broadcast-in-the-international-radio

(2015). Rescue Robotics: An Introduction. iRevolutions, 10 August. Available from https://