**3. Technical aspects**

### **3.1 How technologies can contribute to disaster prevention and critical events notification**

Disasters are caused by nature and/or by human activities, and they can impact on security, agriculture, industry, health, and natural environment. Technologies can have a key role in disaster prevention and management. While some tools, as for example, some environmental sensors, have been on the market since the first half of the twentieth century; in the last 10 years, the opportunity to connect devices through the Internet has created many new opportunities. People can use recent technologies for real-time analytics, remote or in-site monitoring, data analysis,

early warning notification, victim localisation, and data aggregation. Research institutes are the organisations that commonly collect data and try to build some models useful for events forecasting. Those institutes can build their own data acquisition systems, and they can use data collected by others, for example, weather forecast by military or a mix of both. The diffusion of open data databases that people can consult for free on the Internet can help to have more data, but, at the same time, it is important to investigate about how those data were collected and how accurate are them.

Useful raw data can be collected not just from sensor networks but also thanks to IoT devices [7]. If we search information about disaster prevention and IoT, we can find that there is a lot of confusion in the use of the word IoT; unfortunately, this is true also in scientific papers. The biggest confusion is between Internet of Things and sensor networks. There are many definitions of IoT, and we refer to the one made by the Institute of Electric and Electronic Engineers. The document "Towards a definition of the Internet of Things (IoT)" [8] revision 1 (2015) tries to define the concept in 86 pages. We can find that the main concepts, taken from Section 5.3, are the following:


The first point is important: the physical object itself must be relevant from a user or in the application perspective. For example, a temperature sensor that sends data to a remote dashboard is not an IoT device. On the other hand, a fridge that sense that the milk is ended and it informs as with a push notification or an alarm clock that changes the alarm time according to the traffic information are examples of IoT devices.

**Figure 3** shows how data can be collected and used. Data sources can be sensors, other databases, or people notifications. In the last case, people can contribute in many ways, for example, using a mobile application to notify something.

Once raw data are acquired, the business logic can notify alarms or create new data, for example, forecasts, combining trends, and different data sources. The intensity of a measured variable or the result of some statistical elaborations can assume different sense in the phase of sensing, data analysis, and communication to the stakeholders.

**101**

ones (https://models.slf.ch/).

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

**3.2 Environmental risk and sensors and forecasting technique examples**

the market that allow to predict or to early warning disasters.

warning, or for both the scopes together.

before the sending nodes probably die [10-13].

*3.2.1 Landslides*

**Figure 3.**

*Data input and output.*

*3.2.2 Avalanches*

Environmental data are at the base of natural disaster prevention. In this section, we will explore what are some of the best sensor-based solutions currently on

The state of the art for landslide monitoring uses inclinometers, extensometers, and acceleration and piezometer sensors. Before using sensors, it is important to have an inventory with the zones of interest. This can be done to detect landslide that already happened with high-resolution aerial images [9] or LiDAR data. These data must be aggregated to the information that researchers can obtain from city registers, people interviews, other researcher's databases, etc. Once the areas of interest are defined, it is possible to use sensors for advance warning, for event

Inclinometer (see **Figure 4**) and extensometer can be used for advance warning, detecting the small changes in the distance between two points. The most common idea [10–13] for landslide monitoring during the event is to put a lot of nodes on the area of interest; in this way, the signal can be forwarded to the neighbour nodes

Compared to other natural disasters, the risk of avalanches is more to be controlled by humans. Many studies correlate the temperature changes, the type of the snow, and the probability that an avalanche occurs. Humans can also create new controlled avalanches to avoid a disaster, for example, they can trigger some small avalanches before too much snow accumulates on the mountains over a part of a ski resort or a street. In any case, human's understanding of the dynamic response of the snowpack to avalanche control explosives is rather limited [14]. The historical and present weather conditions can be used to forecast the snow conditions in the future. In Europe, the model COSMO (http://www. cosmo-model.org/) is an example of a project developed to create an atmospheric forecasting model. Ref. [15] uses this model in conjunction with a snow cover model to forecast avalanche activity; particularly, they concentrate on wet-snow instability. The WSL Institute for Snow and Avalanche Research SLF creates a repository that includes the snow cover model that they used and other similar

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

*The Internet of Things for Natural Risk Management (Inte.Ri.M.) DOI: http://dx.doi.org/10.5772/intechopen.81707*

**Figure 3.** *Data input and output.*

*Perspectives on Risk, Assessment and Management Paradigms*

how accurate are them.

the following:

protocols.

of IoT devices.

the stakeholders.

application perspective.

early warning notification, victim localisation, and data aggregation. Research institutes are the organisations that commonly collect data and try to build some models useful for events forecasting. Those institutes can build their own data acquisition systems, and they can use data collected by others, for example, weather forecast by military or a mix of both. The diffusion of open data databases that people can consult for free on the Internet can help to have more data, but, at the same time, it is important to investigate about how those data were collected and

Useful raw data can be collected not just from sensor networks but also thanks to IoT devices [7]. If we search information about disaster prevention and IoT, we can find that there is a lot of confusion in the use of the word IoT; unfortunately, this is true also in scientific papers. The biggest confusion is between Internet of Things and sensor networks. There are many definitions of IoT, and we refer to the one made by the Institute of Electric and Electronic Engineers. The document "Towards a definition of the Internet of Things (IoT)" [8] revision 1 (2015) tries to define the concept in 86 pages. We can find that the main concepts, taken from Section 5.3, are

• **Interconnection of Things:** The first feature of IoT is derived from the name that describes it. It is a system that deals with the interconnection of "Things". The word "Thing" refers to any physical object that is relevant from a user or

• **Connection of Things to the Internet:** From the name IoT, we can also learn that the "Things" are connected to the Internet. Accordingly, from the name we can deduce that the system is not an Intranet or Extranet of Things.

• **Sensing/actuation capability:** There is the involvement of sensors/actuators in the IoT system. The sensors/actuators are connected to the "Things" and perform the sensing and actuation, which bring the smartness of the "Things".

• **Embedded intelligence:** Smart and dynamic objects, with emergent behaviour, embed intelligence and knowledge functions as tools and become an

• **Interoperable communication capability:** The IoT system has a com-

munication capability based on standard and interoperable communication

The first point is important: the physical object itself must be relevant from a user or in the application perspective. For example, a temperature sensor that sends data to a remote dashboard is not an IoT device. On the other hand, a fridge that sense that the milk is ended and it informs as with a push notification or an alarm clock that changes the alarm time according to the traffic information are examples

**Figure 3** shows how data can be collected and used. Data sources can be sensors, other databases, or people notifications. In the last case, people can contribute in

Once raw data are acquired, the business logic can notify alarms or create new data, for example, forecasts, combining trends, and different data sources. The intensity of a measured variable or the result of some statistical elaborations can assume different sense in the phase of sensing, data analysis, and communication to

many ways, for example, using a mobile application to notify something.

(external) extension to the human body and mind.

**100**

#### **3.2 Environmental risk and sensors and forecasting technique examples**

Environmental data are at the base of natural disaster prevention. In this section, we will explore what are some of the best sensor-based solutions currently on the market that allow to predict or to early warning disasters.

#### *3.2.1 Landslides*

The state of the art for landslide monitoring uses inclinometers, extensometers, and acceleration and piezometer sensors. Before using sensors, it is important to have an inventory with the zones of interest. This can be done to detect landslide that already happened with high-resolution aerial images [9] or LiDAR data. These data must be aggregated to the information that researchers can obtain from city registers, people interviews, other researcher's databases, etc. Once the areas of interest are defined, it is possible to use sensors for advance warning, for event warning, or for both the scopes together.

Inclinometer (see **Figure 4**) and extensometer can be used for advance warning, detecting the small changes in the distance between two points. The most common idea [10–13] for landslide monitoring during the event is to put a lot of nodes on the area of interest; in this way, the signal can be forwarded to the neighbour nodes before the sending nodes probably die [10-13].

#### *3.2.2 Avalanches*

Compared to other natural disasters, the risk of avalanches is more to be controlled by humans. Many studies correlate the temperature changes, the type of the snow, and the probability that an avalanche occurs. Humans can also create new controlled avalanches to avoid a disaster, for example, they can trigger some small avalanches before too much snow accumulates on the mountains over a part of a ski resort or a street. In any case, human's understanding of the dynamic response of the snowpack to avalanche control explosives is rather limited [14].

The historical and present weather conditions can be used to forecast the snow conditions in the future. In Europe, the model COSMO (http://www. cosmo-model.org/) is an example of a project developed to create an atmospheric forecasting model. Ref. [15] uses this model in conjunction with a snow cover model to forecast avalanche activity; particularly, they concentrate on wet-snow instability. The WSL Institute for Snow and Avalanche Research SLF creates a repository that includes the snow cover model that they used and other similar ones (https://models.slf.ch/).

**Figure 4.** *Inclinometer probes allow to detect ground movements under the soil.*

If the historic trend of environmental parameter such as temperature and snow level is really important to discover the avalanche's probability, in situ test made by people can be also quite important to discover how much the layers of snow are discontinuous.

For avalanche monitoring, the most common sensor solutions are **ultrasound sensors**, **Doppler radars**, and **optic sensors** [16]. Other solutions include **cameras** and image recognition, LiDAR, or the use of **geophone** built for earthquake monitoring [17]. These seismic sensors have two roles: monitoring the activity of the avalanche and studying the dynamics of the avalanche. It is interesting to notice that every avalanche does not only produce a seismic signal but also some low frequency infrasonic acoustic waves (below 20 Hz). Remote infrasound solutions can be used for monitoring also when the visibility is low [18].

Using geophones, it is important to discriminate signals due to the avalanche from other noises. The analysis in the time-frequency domain allows to distinguish different sources, for example, helicopters or airplanes that generate signals on some specific harmonic frequencies [19]. It is also interesting to notice that diverse kinds of avalanche; for example, loose snow and slab ones have their typical signal characteristics. Detecting precursor signal with geophones is very difficult but these instruments allow not just to detect avalanches but also to study the dynamic of the snow inside the avalanche.

Another interesting aspect of snow to be monitored is the snow flow. The snow can be moved by the wind, and in this way, it can influence the avalanche danger, it can block some roads or the access to some buildings, and it can cause other problems. Lehning and others, in the article *Snow drift: acoustic sensors for avalanche warning and research*, compare many kinds of sensor for snow flow; in particular, they concentrate their analysis on the acoustics sensors. Their findings are that, at the present, it is not possible to measure the mass of the flux of snow with high accuracy [20]. A significant reduction of false detection in infrasonic measure can be obtained using machine learning classification techniques compared to the use of a threshold [16].

One problem of snow monitoring is also that, in all the previous cases, you can only monitor a specific area, but the snow condition can be quite different in another place not far from the installed sensors.

New technologies have an important role not just in avalanche detection but also for localising the victims. Common instruments like the ARTVA allow to find people under the avalanche. New tools like wearable sensors are currently not so diffused, but in future, they will probably help the rescue operations, giving information about the health status of the victims under the snow [21].

**103**

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

disaster in cities must also take into account the sewer capacity [25].

learning techniques for predicting values can also help.

The risk of flood disaster increased in the last years because people build more buildings on floodplains, and the area covered by cement has increased. To make actions against flood effective, it is important to forecast it further in advance as possible. Usually people working in the emergency response staff are notified before common people. This happen because early previsions are more subject to false alarm [22]. The predictability can change a lot depending on the historical data and the place. For example, Webster et al. in their article demonstrate that it was possible to predict the 2010 floods in Pakistan 6–8 days in advance using a multiyear analysis and a hydrological model to mitigate the impact anticipating actions [23]. Rain sensors alone cannot forecast floods; the most important prevision tools are software using mathematical algorithms based upon time series. The data for the analysis can be a mix among rain gauge measures and satellite information. Prediction models useful for short-range prediction can differ from the ones that are best for long-term forecasts [24]. It is interesting to notice that simulate flood

Other types of events to be forecasted for safety reasons are the wave run-up and the wave overtopping. The project HIDRALERTA [26] developed for the safeguard of the Portugal coasts is an interesting case because it is a tool for evaluating the risk, taking a decision, and eventually alerting the involved stakeholders. The explained methodology is composed by four parts and can be applied to different

• The wave characterisation: This step can be accomplished with some spectral or mid slope wave models. These predictive models need the sea to be monitored and, being too much complex, can be simplified using some matrix with some pre-run different models for multiple conditions. The use of machine

• A neural network determines if the wave will overtop the barrier or not.

• Risk level calculation for the overtopping values. Risk maps are produced taking into account the thresholds for the overlapping, the probability for the

As for floods, also for earthquakes, the most important prevision tools are software using mathematical algorithms based upon time series. Earthquakes can have a bigger impact if they affect some human structures like a dike or a nuclear power plant. Herle and others [27] simulate many dangerous events for calculating the risk of a dike failure. In their simulation, they use a protocol called message queue telemetry transport (MQTT) that is a lightweight IoT transmission protocol. We can find an interesting approach in the article by Zambrano et al. [28]. They build a system for early detecting and notifying earthquakes using people smartphones' accelerometers. The work is a notable example of a project that can be developed almost without having new costs and using existing devices for a new scope. Munib ur Rehman and others in their article analyse how ICT and wireless sensor

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

geographic area with different data sources:

event to occur, and the consequences.

• The warning and alarm system.

*3.2.4 Earthquakes*

*3.2.3 Floods*

*The Internet of Things for Natural Risk Management (Inte.Ri.M.) DOI: http://dx.doi.org/10.5772/intechopen.81707*

#### *3.2.3 Floods*

*Perspectives on Risk, Assessment and Management Paradigms*

*Inclinometer probes allow to detect ground movements under the soil.*

If the historic trend of environmental parameter such as temperature and snow level is really important to discover the avalanche's probability, in situ test made by people can be also quite important to discover how much the layers of snow are

For avalanche monitoring, the most common sensor solutions are **ultrasound sensors**, **Doppler radars**, and **optic sensors** [16]. Other solutions include **cameras** and image recognition, LiDAR, or the use of **geophone** built for earthquake monitoring [17]. These seismic sensors have two roles: monitoring the activity of the avalanche and studying the dynamics of the avalanche. It is interesting to notice that every avalanche does not only produce a seismic signal but also some low frequency infrasonic acoustic waves (below 20 Hz). Remote infrasound solutions

Using geophones, it is important to discriminate signals due to the avalanche from other noises. The analysis in the time-frequency domain allows to distinguish different sources, for example, helicopters or airplanes that generate signals on some specific harmonic frequencies [19]. It is also interesting to notice that diverse kinds of avalanche; for example, loose snow and slab ones have their typical signal characteristics. Detecting precursor signal with geophones is very difficult but these instruments allow not just to detect avalanches but also to study the dynamic of the

Another interesting aspect of snow to be monitored is the snow flow. The snow can be moved by the wind, and in this way, it can influence the avalanche danger, it can block some roads or the access to some buildings, and it can cause other problems. Lehning and others, in the article *Snow drift: acoustic sensors for avalanche warning and research*, compare many kinds of sensor for snow flow; in particular, they concentrate their analysis on the acoustics sensors. Their findings are that, at the present, it is not possible to measure the mass of the flux of snow with high accuracy [20]. A significant reduction of false detection in infrasonic measure can be obtained using machine learning classification techniques compared to the use of

One problem of snow monitoring is also that, in all the previous cases, you can only monitor a specific area, but the snow condition can be quite different in

New technologies have an important role not just in avalanche detection but also for localising the victims. Common instruments like the ARTVA allow to find people under the avalanche. New tools like wearable sensors are currently not so diffused, but in future, they will probably help the rescue operations, giving

information about the health status of the victims under the snow [21].

can be used for monitoring also when the visibility is low [18].

**102**

discontinuous.

**Figure 4.**

snow inside the avalanche.

a threshold [16].

another place not far from the installed sensors.

The risk of flood disaster increased in the last years because people build more buildings on floodplains, and the area covered by cement has increased. To make actions against flood effective, it is important to forecast it further in advance as possible. Usually people working in the emergency response staff are notified before common people. This happen because early previsions are more subject to false alarm [22]. The predictability can change a lot depending on the historical data and the place. For example, Webster et al. in their article demonstrate that it was possible to predict the 2010 floods in Pakistan 6–8 days in advance using a multiyear analysis and a hydrological model to mitigate the impact anticipating actions [23]. Rain sensors alone cannot forecast floods; the most important prevision tools are software using mathematical algorithms based upon time series. The data for the analysis can be a mix among rain gauge measures and satellite information. Prediction models useful for short-range prediction can differ from the ones that are best for long-term forecasts [24]. It is interesting to notice that simulate flood disaster in cities must also take into account the sewer capacity [25].

Other types of events to be forecasted for safety reasons are the wave run-up and the wave overtopping. The project HIDRALERTA [26] developed for the safeguard of the Portugal coasts is an interesting case because it is a tool for evaluating the risk, taking a decision, and eventually alerting the involved stakeholders. The explained methodology is composed by four parts and can be applied to different geographic area with different data sources:


#### *3.2.4 Earthquakes*

As for floods, also for earthquakes, the most important prevision tools are software using mathematical algorithms based upon time series. Earthquakes can have a bigger impact if they affect some human structures like a dike or a nuclear power plant. Herle and others [27] simulate many dangerous events for calculating the risk of a dike failure. In their simulation, they use a protocol called message queue telemetry transport (MQTT) that is a lightweight IoT transmission protocol. We can find an interesting approach in the article by Zambrano et al. [28]. They build a system for early detecting and notifying earthquakes using people smartphones' accelerometers. The work is a notable example of a project that can be developed almost without having new costs and using existing devices for a new scope. Munib ur Rehman and others in their article analyse how ICT and wireless sensor

networks can contribute to detect earthquakes early. They used three main different approaches correlating three events with the possibility of an earthquake. Detecting abnormal animals' movement can help to early prevent calamity, and also variation in groundwater pressure gives information about the possibility of earthquakes. Another precursor is the concentration of radon on the ground that changes because of the breakings of rocks below the surface.

Sensors can give an important contribution also after a disaster; for example, they can monitor pollutants and contaminants including radioactive scenarios after an earthquake has damaged an industrial area.

#### *3.2.5 Debris flows*

Another dangerous process that can occur during strong rainfalls is the debris flow. Compared to landslides, the debris flows are more difficult to study because these events can occur in a very short time, and using aerial photographs, it is hard to find traces of previous events [29]. An approach for forecasting debris flows can be the creation of a map with the alluvial fan using LiDAR data and photograph and keeping trace of the event occurrences. Because this event can be considered as the union of landslides and water floods, sensors can help to detect strong waterfall and notify early warning. Example of sensors for this application is ultrasonic or radar gauges, ground vibration sensors, video cameras, avalanche pendulums, photocells, and trip wires [30].

#### *3.2.6 Fire*

It is very hard to forecast a fire. It can be caused directly by humans or by some objects like trains or electrical lines. If we want to limit fire consequences, it is important to detect the fire as soon as possible. A possible approach is to use IP cameras to detect the smoke at its beginning using some detection algorithms [31, 32, 33]. Another approach is to use geostationary satellites and image processing [34].

Fire can be early detected even before the smoke presence using gas sensor as demonstrated by Gutmacher et al. [35] in the article "Gas Sensor Technologies for Fire Detection". In their experiment, the fire was detected by gas sensors 4 minutes before one optic smoke sensor detected it. This approach can be used mostly in closed areas, for example, offices and houses. In the same context, also heat detectors can provide early earning of fire incidents.

An interesting experiment in outdoor context [36] uses the wireless sensor network transmission signal strength indicator variance between two nodes to detect the fire. The idea is that the changes in the vegetation water content level due to the fire influence the wireless signal.

#### **3.3 Other scenarios and solutions**

Human activities can have an appreciable impact on natural disaster activities can have as a consequence relevant natural disasters, for example, the Chernobyl nuclear disaster that happened during some scheduled test. Also, in the Seveso disaster, people working in the company had many responsibilities [37]. Apart from human dangerous behaviours and errors, when an organisation is conducting an environmentally risky activity, it is important to continuously monitor the relevant processes.

Some very diffused technologies can help during disasters: smartphone, tablet, and other devices are useful for the alert and the management steps of emergencies. They can also collect data from people in the emergency area, for example,

**105**

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

the device can share the position of an injured person using the GPS. Also, social network like Twitter and Facebook can help to collect data from people in the zone

Lapante and others in the article "Could the Internet of Things Be Used to Enhance Student Nurses' Experiences in a Disaster Simulation?" [38] illustrate the simulation of a street accident with many people injured. Nurses and patients in the place are connected and can share information about the patients, thanks to some wearable sensors and the use of RFID for patient recognition. If their approach is not realistic because people are not usually equipped with a RFID system for recognition, an interesting approach is the one of Libelium Meshlium. This product, which is currently on the market, allows to count people in a space looking for the Bluetooth and Wi-Fi signal of their mobile phones. This approach cannot assure to count the total amount of persons, but it is easy, and it is applicable in real word and

Many technologies are useful for collecting data from the sensor nodes in a network. **Table 1** shows some of the currently used communication technologies. In the most common scenario, the Internet network is used to transmit data to a remote server. It is important to underline that different protocols can operate together: for example, we can have one protocol for data collection among the nodes of a wireless sensor network (WSN) and another one from one collector node in the

The data communication is easier if we collect data for doing forecast but is more critical if we need to transmit information in real time during the disaster. The main problems in transmitting data in such conditions are the availability of power supply for the nodes and of the data network for transmitting the data packets. Many research reports (e.g., [39]) discuss the importance of redundancy. It is possible to use some battery to assure that the node will transmit data also if there is no external power, but the most critical point happens if the data network is absent. For example, if the cellular base transceiver station of our operator is not available, it becomes important to use the one of another operator. All the mobile phones permit to make emergency call independently from the operator and the credit, but

In some applications, for example, avalanche monitoring, we can put many radios to have more chances that almost one radio will successfully transmit. The adoption of a mesh network, where a packet can take many paths to reach its

The importance of having a working system during the disaster depends on the specific application. For example, at that time, information needed for prevention can be less useful with respect to data useful to localise injured people overwhelmed during an earthquake. The network services can be unavailable not just because the radios are physically broken but also because there are too many requests: for example, many people trying to call with their mobile phones just after the disaster

Before proceeding with the installation of the monitoring network prototype, it is considered useful to test the performance of the network in a mountain environment. For this reason, the site of the Angelo Mosso Institute of the University of Turin has been chosen. It hosts groups of international researchers working in various scientific areas. The objective of this phase is the verification of the energy performance of the remote sensing network in relation to the frequency of data

this is not the same for other services like sharing the position.

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

**3.4 Possible issues in data communication**

of the emergency.

not just in simulations.

WSN and the Internet.

destination, also helps.

can saturate the cell.

transmission [40].

*The Internet of Things for Natural Risk Management (Inte.Ri.M.) DOI: http://dx.doi.org/10.5772/intechopen.81707*

*Perspectives on Risk, Assessment and Management Paradigms*

of the breakings of rocks below the surface.

an earthquake has damaged an industrial area.

tors can provide early earning of fire incidents.

fire influence the wireless signal.

**3.3 Other scenarios and solutions**

*3.2.5 Debris flows*

and trip wires [30].

*3.2.6 Fire*

networks can contribute to detect earthquakes early. They used three main different approaches correlating three events with the possibility of an earthquake. Detecting abnormal animals' movement can help to early prevent calamity, and also variation in groundwater pressure gives information about the possibility of earthquakes. Another precursor is the concentration of radon on the ground that changes because

Sensors can give an important contribution also after a disaster; for example, they can monitor pollutants and contaminants including radioactive scenarios after

Another dangerous process that can occur during strong rainfalls is the debris flow. Compared to landslides, the debris flows are more difficult to study because these events can occur in a very short time, and using aerial photographs, it is hard to find traces of previous events [29]. An approach for forecasting debris flows can be the creation of a map with the alluvial fan using LiDAR data and photograph and keeping trace of the event occurrences. Because this event can be considered as the union of landslides and water floods, sensors can help to detect strong waterfall and notify early warning. Example of sensors for this application is ultrasonic or radar gauges, ground vibration sensors, video cameras, avalanche pendulums, photocells,

It is very hard to forecast a fire. It can be caused directly by humans or by some objects like trains or electrical lines. If we want to limit fire consequences, it is important to detect the fire as soon as possible. A possible approach is to use IP cameras to detect the smoke at its beginning using some detection algorithms [31, 32, 33]. Another approach is to use geostationary satellites and image processing [34]. Fire can be early detected even before the smoke presence using gas sensor as demonstrated by Gutmacher et al. [35] in the article "Gas Sensor Technologies for Fire Detection". In their experiment, the fire was detected by gas sensors 4 minutes before one optic smoke sensor detected it. This approach can be used mostly in closed areas, for example, offices and houses. In the same context, also heat detec-

An interesting experiment in outdoor context [36] uses the wireless sensor network transmission signal strength indicator variance between two nodes to detect the fire. The idea is that the changes in the vegetation water content level due to the

Human activities can have an appreciable impact on natural disaster activities can have as a consequence relevant natural disasters, for example, the Chernobyl nuclear disaster that happened during some scheduled test. Also, in the Seveso disaster, people working in the company had many responsibilities [37]. Apart from human dangerous behaviours and errors, when an organisation is conducting an environmentally risky activity, it is important to continuously monitor the relevant

Some very diffused technologies can help during disasters: smartphone, tablet, and other devices are useful for the alert and the management steps of emergencies. They can also collect data from people in the emergency area, for example,

**104**

processes.

the device can share the position of an injured person using the GPS. Also, social network like Twitter and Facebook can help to collect data from people in the zone of the emergency.

Lapante and others in the article "Could the Internet of Things Be Used to Enhance Student Nurses' Experiences in a Disaster Simulation?" [38] illustrate the simulation of a street accident with many people injured. Nurses and patients in the place are connected and can share information about the patients, thanks to some wearable sensors and the use of RFID for patient recognition. If their approach is not realistic because people are not usually equipped with a RFID system for recognition, an interesting approach is the one of Libelium Meshlium. This product, which is currently on the market, allows to count people in a space looking for the Bluetooth and Wi-Fi signal of their mobile phones. This approach cannot assure to count the total amount of persons, but it is easy, and it is applicable in real word and not just in simulations.

#### **3.4 Possible issues in data communication**

Many technologies are useful for collecting data from the sensor nodes in a network. **Table 1** shows some of the currently used communication technologies. In the most common scenario, the Internet network is used to transmit data to a remote server. It is important to underline that different protocols can operate together: for example, we can have one protocol for data collection among the nodes of a wireless sensor network (WSN) and another one from one collector node in the WSN and the Internet.

The data communication is easier if we collect data for doing forecast but is more critical if we need to transmit information in real time during the disaster. The main problems in transmitting data in such conditions are the availability of power supply for the nodes and of the data network for transmitting the data packets.

Many research reports (e.g., [39]) discuss the importance of redundancy. It is possible to use some battery to assure that the node will transmit data also if there is no external power, but the most critical point happens if the data network is absent. For example, if the cellular base transceiver station of our operator is not available, it becomes important to use the one of another operator. All the mobile phones permit to make emergency call independently from the operator and the credit, but this is not the same for other services like sharing the position.

In some applications, for example, avalanche monitoring, we can put many radios to have more chances that almost one radio will successfully transmit. The adoption of a mesh network, where a packet can take many paths to reach its destination, also helps.

The importance of having a working system during the disaster depends on the specific application. For example, at that time, information needed for prevention can be less useful with respect to data useful to localise injured people overwhelmed during an earthquake. The network services can be unavailable not just because the radios are physically broken but also because there are too many requests: for example, many people trying to call with their mobile phones just after the disaster can saturate the cell.

Before proceeding with the installation of the monitoring network prototype, it is considered useful to test the performance of the network in a mountain environment. For this reason, the site of the Angelo Mosso Institute of the University of Turin has been chosen. It hosts groups of international researchers working in various scientific areas. The objective of this phase is the verification of the energy performance of the remote sensing network in relation to the frequency of data transmission [40].


#### **Table 1.**

*Technologies suitable for data transmission in disaster scenario.*

#### **4. Economic aspects**

The economic part of the Inte.Ri.M project aims at selecting correct methodologies to evaluate and to allow the comparison among costs of prevention and the damages from the disasters that can be avoided. In order to achieve this ambitious goal, an overall strategy is defined, and three parts are identified. The first is about the different scenario to analyse, depending of the type of territory that disaster may affect, and the second is related to the different methodologies used for evaluation. The third part is focused on the actual possibility to compare the costs and benefits of disaster prevention.

#### **4.1 Territory and disasters**

According to the area that disasters affect, it is possible to identify different situations. Inte.Ri.M. project has focused its attention on mountain ecosystem (Piedmont region) and above all on specific natural disaster: avalanches and landslides, with circumscribed and localised impact. As seen, the same reasoning can be

**107**

**Figure 5.**

*Presence of ecosystem or assets economically evaluables.*

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

extended to other geophysical (like earthquakes and forest fires) and hydrological

from the consequences. Both are important for economic evaluations.

For the purposes of ongoing research, two classification parameters have been identified: the *type of territory* affected by the disaster and the time taken to recover

The first, it is fundamental to choose the correct variables to be considered. It is very different to destroy an urban area from a completely uninhabited area. Thus, as a sample area, two different types were chosen, always in a mountain environment:

In **Figure 5**, the presence of the ecosystem services (ESS) or economic compo-

The second, it is useful to evaluate the "time axis" that sets the economic reasoning. **Table 2** shows the various recovery times for different disaster and for type of

This subdivision based on the time horizons allows, within the economic assessment, to make a more correct comparison with the benefits of the avoided damage. To consider that the comparisons must be homogeneous also from the temporal point of view with the benefits that derive from the systems that disasters are going to disturb. In the case of ecosystems, as well as in traditional economic systems, the perspective is that of the medium/long term (Ministero dell'ambiente e della tutela

First of all, we identified what are the information concerning natural disasters to be taken into consideration. After that, we proceed to build a scheme useful for

To achieve this, it is necessary to hypothesise which valuation objects would have been damaged and, on the basis of this, verify which tools are available to give an economic value. Ecosystems played an important role in this process. Even if they can be classified into natural, semi-natural, and artificial [42], for the natural territories, the presence of natural ecosystems was particularly assumed, and therefore, we have taken into consideration the damage only to this system. Therefore, the urbanised areas are not considered as artificial ecosystems (urban ecosystem), and other methodologies are used to calculate damages on people, infrastructure, and

**4.2 Disaster on highly natural area: the value of ecosystem service (ESS)**

When it was agreed that economy and ecology had to go in tandem (cfr "ecological economy") against a background of unsustainable growth in the resources use, common language has been sought. The first step to solve this challenge is to refer to a common functional unit: the ecosystem (Ministero dell'ambiente e della tutela del

nents can be noticed, depending on the type of area taken into consideration.

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

a highly natural area and an urbanised area.

del territorio e del mare, 2009).

economic evaluation.

industrial plants.

disasters [41].

area.

*The Internet of Things for Natural Risk Management (Inte.Ri.M.) DOI: http://dx.doi.org/10.5772/intechopen.81707*

*Perspectives on Risk, Assessment and Management Paradigms*

**topology**

users view)

Fully connected star

star (mesh supported)

star

*Technologies suitable for data transmission in disaster scenario.*

**Average ranges Average Tx** 

ZigBee (s2/s2pro) Mesh 20 m/3.2 km 3.1 mW/63 mW Many nodes, node

xBee s5 868/900 Star 500 m–12 km 315 mW Have to collect data in

Sigfox Star 30 km 300 mW Direct communication to

Satellite ptp 36,000 km 2–5 W There are no other ways

Bluetooth star 50 m 10 mW Need to interface with a

Typical carrier ranges 30 km

**power**

120 m/3.2 km 10 mW Low power, not many

5–15 km 100 mW Need to transmit small

50 m 100 mW Can use an already built

**Scenario**

routing needed

a remote server thought the Internet. No LAN available

nodes

open spaces. Do not need packet routing

amount of data covering distances in Km order. Need a WAN gateway in the same technology as the nodes

a remote server thought Internet, less expensive than 3G. Only few access points available

to connect. For example, you are in the sea or in a desert

Wi-Fi network. Do not need packet routing

Bluetooth device

2 W Direct communication to

**Name Network** 

3G/4G ptp (from

Lora Typically

Wi-Fi Typically

xBee 802.15.4 (s1/ s1 pro)

**106**

**4. Economic aspects**

**Table 1.**

benefits of disaster prevention.

**4.1 Territory and disasters**

The economic part of the Inte.Ri.M project aims at selecting correct methodologies to evaluate and to allow the comparison among costs of prevention and the damages from the disasters that can be avoided. In order to achieve this ambitious goal, an overall strategy is defined, and three parts are identified. The first is about the different scenario to analyse, depending of the type of territory that disaster may affect, and the second is related to the different methodologies used for evaluation. The third part is focused on the actual possibility to compare the costs and

According to the area that disasters affect, it is possible to identify different situations. Inte.Ri.M. project has focused its attention on mountain ecosystem (Piedmont region) and above all on specific natural disaster: avalanches and landslides, with circumscribed and localised impact. As seen, the same reasoning can be extended to other geophysical (like earthquakes and forest fires) and hydrological disasters [41].

For the purposes of ongoing research, two classification parameters have been identified: the *type of territory* affected by the disaster and the time taken to recover from the consequences. Both are important for economic evaluations.

The first, it is fundamental to choose the correct variables to be considered. It is very different to destroy an urban area from a completely uninhabited area. Thus, as a sample area, two different types were chosen, always in a mountain environment: a highly natural area and an urbanised area.

In **Figure 5**, the presence of the ecosystem services (ESS) or economic components can be noticed, depending on the type of area taken into consideration.

The second, it is useful to evaluate the "time axis" that sets the economic reasoning. **Table 2** shows the various recovery times for different disaster and for type of area.

This subdivision based on the time horizons allows, within the economic assessment, to make a more correct comparison with the benefits of the avoided damage. To consider that the comparisons must be homogeneous also from the temporal point of view with the benefits that derive from the systems that disasters are going to disturb. In the case of ecosystems, as well as in traditional economic systems, the perspective is that of the medium/long term (Ministero dell'ambiente e della tutela del territorio e del mare, 2009).

First of all, we identified what are the information concerning natural disasters to be taken into consideration. After that, we proceed to build a scheme useful for economic evaluation.

To achieve this, it is necessary to hypothesise which valuation objects would have been damaged and, on the basis of this, verify which tools are available to give an economic value. Ecosystems played an important role in this process. Even if they can be classified into natural, semi-natural, and artificial [42], for the natural territories, the presence of natural ecosystems was particularly assumed, and therefore, we have taken into consideration the damage only to this system. Therefore, the urbanised areas are not considered as artificial ecosystems (urban ecosystem), and other methodologies are used to calculate damages on people, infrastructure, and industrial plants.

#### **4.2 Disaster on highly natural area: the value of ecosystem service (ESS)**

When it was agreed that economy and ecology had to go in tandem (cfr "ecological economy") against a background of unsustainable growth in the resources use, common language has been sought. The first step to solve this challenge is to refer to a common functional unit: the ecosystem (Ministero dell'ambiente e della tutela del

**Figure 5.** *Presence of ecosystem or assets economically evaluables.*


#### **Table 2.**

*Time axis for disasters and areas of different types.*

territorio e del mare, 2009). Because it is very difficult to evaluate an ecosystem for the multiple relationships inside, we value ESS to human well-being [43].

This aspect gives us a chance on developing a different logical argument. In fact, rarely, the discussion on an area's value is focused on ESS: normally all attention is on the role that plays on local economies. Areas have to be recognised for the service they provide: mitigation of climate change and natural disasters, disease control, maintenance of water quality, and cultural services, including recreation, maintenance of historical or iconic landscapes, and protection of sacred natural sites [44].

ESSs are defined by the Millennium Ecosystem Assessment as "the benefits people obtain from ecosystems" [3]. Fisher and Turner [45] added that "ecosystems services are the aspects of ecosystems utilised (actively or passively) to produce human well-being". The MA developed and claimed a very important concept: the well-being of society is strictly linked to services provided by nature. **Figure 6**, by MA [3], shows this relation.

The measurement of losses' full costs and degradation of these ESSs are difficult to assess, but there is a sense of urgency to implement an economic tracking system to monitor all disaster risk reduction costs for mitigation, preparedness, and emergency response. A correct evaluation of ESS gives helpful information to estimate costs and benefits of decisions, to define future scenarios and to recognise and avoid unexpected consequences [5]. The value of ESS can be linked to the direct use (e.g., wood) or indirect use (value of specific ecological functions) or to the particular item relevance for other goods or values, or, finally, the value may simply be intrinsic (e.g., existence value or cultural value). To narrow the field and make a good comparison, the first step is to draw from the studies that have circumscribed key ESS of the different biomes on Earth [3]. The starting point is the ecosystems present in the investigated areas. The logical path is that of analysing the identification of the services offered, to move on to quantification and valorisation [46]. In this case, we consider ESS in alpine environments (mountains and forests), which mainly serve areas with high naturalness but also urbanised areas, always in a mountain environment. For the *identification* step, it is important to start from ESS definition. If ESSs are useful for human well-being, the supply of and demand for ESS have to be determined [47]. In order to define which type of ESS the territory can supply, the reviews cited by Haida et al. [48] have reported that most investigated are regulating services (mainly carbon sequestration) followed by provisioning services (such as the supply of fresh water and food) and cultural services (largely recreation and tourism). From the demand side, an analysis of inhabitants, touristic flows, and activities must be conducted (on bibliographic basis). The steps of *quantification* and *evaluation* show the main difficulties to arrive at a concrete and usable result. Starting from Costanza's study, in one of the most famous ecological economists, where dozens of scientific papers on economic evaluation of ESS [43] are summarised [43] and produce a "summary of average global value of annual ecosystem services", many other researches have tried to get into the specifics of this complex assessment.

**109**

development policies.

*Ecosystem services and human well-being.*

**Figure 6.**

different ESSs.

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

Despite all the progresses made in this field, it is difficult to assess environmental assets, especially when it is necessary to attribute value to the benefits they generate. Even if the monetary evaluation of environmental goods without a market can be more or less correct, it provides a starting point for hypothesising policies and strategies. In order to focus attention and attribute a monetary value to environmental goods and, in this case, to ESS, the willingness to pay (WtP) and total economic value (VET) are often used. Also, in his review reports, WTP of individuals for ESS is the first method used [43]. In this case, it is emphasised that it is not a matter of attributing a value to a real exchange but only to make comparisons and establish strategies. Monetary value can be used as an indicator to encourage

Instead, as in the case of PES (payment for ecosystem services) [46], governments and private enterprises have started to pay for ESS [44]. Other authors have stated that its price does not represent the good real value but only one way to

For quantification and attribution of value in the development of the project, reference was made to the data sheets deriving from Making Good Natura (LIFE+11 ENV/IT/000168) [5]: within it is possible to find indications for the evaluation of

**4.3 Disaster on urbanised area: the value of life, infrastructures, and plant**

In this part, the urbanised area is taken into consideration, but "urbanised like" can be in mountain territory, the focus of the Inte.RI.M. project. The urban situation is easier to assess but only apparently. It is true that there are methods for evaluating human lives and economic activities, but for these too, there are aspects

facilitate comparison with actions that have a market [49].

of uncertainty as well as those predicted for ESS.

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

*The Internet of Things for Natural Risk Management (Inte.Ri.M.) DOI: http://dx.doi.org/10.5772/intechopen.81707*

*Perspectives on Risk, Assessment and Management Paradigms*

Natural area Earthquakes,

Urbanised area

**Table 2.**

avalanches

*Time axis for disasters and areas of different types.*

territorio e del mare, 2009). Because it is very difficult to evaluate an ecosystem for

**Short term Medium term Long term**

Landslides, avalanches

Landslides Landslides

Earthquakes, landslides,

avalanches

This aspect gives us a chance on developing a different logical argument. In fact, rarely, the discussion on an area's value is focused on ESS: normally all attention is on the role that plays on local economies. Areas have to be recognised for the service they provide: mitigation of climate change and natural disasters, disease control, maintenance of water quality, and cultural services, including recreation, maintenance of historical or iconic landscapes, and protection of sacred natural sites [44]. ESSs are defined by the Millennium Ecosystem Assessment as "the benefits people obtain from ecosystems" [3]. Fisher and Turner [45] added that "ecosystems services are the aspects of ecosystems utilised (actively or passively) to produce human well-being". The MA developed and claimed a very important concept: the well-being of society is strictly linked to services provided by nature. **Figure 6**, by MA [3], shows this relation. The measurement of losses' full costs and degradation of these ESSs are difficult to assess, but there is a sense of urgency to implement an economic tracking system to monitor all disaster risk reduction costs for mitigation, preparedness, and emergency response. A correct evaluation of ESS gives helpful information to estimate costs and benefits of decisions, to define future scenarios and to recognise and avoid unexpected consequences [5]. The value of ESS can be linked to the direct use (e.g., wood) or indirect use (value of specific ecological functions) or to the particular item relevance for other goods or values, or, finally, the value may simply be intrinsic (e.g., existence value or cultural value). To narrow the field and make a good comparison, the first step is to draw from the studies that have circumscribed key ESS of the different biomes on Earth [3]. The starting point is the ecosystems present in the investigated areas. The logical path is that of analysing the identification of the services offered, to move on to quantification and valorisation [46]. In this case, we consider ESS in alpine environments (mountains and forests), which mainly serve areas with high naturalness but also urbanised areas, always in a mountain environment. For the *identification* step, it is important to start from ESS definition. If ESSs are useful for human well-being, the supply of and demand for ESS have to be determined [47]. In order to define which type of ESS the territory can supply, the reviews cited by Haida et al. [48] have reported that most investigated are regulating services (mainly carbon sequestration) followed by provisioning services (such as the supply of fresh water and food) and cultural services (largely recreation and tourism). From the demand side, an analysis of inhabitants, touristic flows, and activities must be conducted (on bibliographic basis). The steps of *quantification* and *evaluation* show the main difficulties to arrive at a concrete and usable result. Starting from Costanza's study, in one of the most famous ecological economists, where dozens of scientific papers on economic evaluation of ESS [43] are summarised [43] and produce a "summary of average global value of annual ecosystem services", many other researches have tried to get into the specifics of this

the multiple relationships inside, we value ESS to human well-being [43].

**108**

complex assessment.

**Figure 6.** *Ecosystem services and human well-being.*

Despite all the progresses made in this field, it is difficult to assess environmental assets, especially when it is necessary to attribute value to the benefits they generate. Even if the monetary evaluation of environmental goods without a market can be more or less correct, it provides a starting point for hypothesising policies and strategies. In order to focus attention and attribute a monetary value to environmental goods and, in this case, to ESS, the willingness to pay (WtP) and total economic value (VET) are often used. Also, in his review reports, WTP of individuals for ESS is the first method used [43]. In this case, it is emphasised that it is not a matter of attributing a value to a real exchange but only to make comparisons and establish strategies. Monetary value can be used as an indicator to encourage development policies.

Instead, as in the case of PES (payment for ecosystem services) [46], governments and private enterprises have started to pay for ESS [44]. Other authors have stated that its price does not represent the good real value but only one way to facilitate comparison with actions that have a market [49].

For quantification and attribution of value in the development of the project, reference was made to the data sheets deriving from Making Good Natura (LIFE+11 ENV/IT/000168) [5]: within it is possible to find indications for the evaluation of different ESSs.

#### **4.3 Disaster on urbanised area: the value of life, infrastructures, and plant**

In this part, the urbanised area is taken into consideration, but "urbanised like" can be in mountain territory, the focus of the Inte.RI.M. project. The urban situation is easier to assess but only apparently. It is true that there are methods for evaluating human lives and economic activities, but for these too, there are aspects of uncertainty as well as those predicted for ESS.

In order to monetise the value of a *human life* in disasters, we often refer to VOSL (value of statistic life) and DALY (disability-adjusted life year) [50].

For a "complete and detailed assessment of the exposure of 'economic activities' to natural disaster", Marin et al. study [51] refers to literature that recommends the use of different proxies depending on type of disaster. For flood, an interesting study carried out on an Italian case exploits the overlap between risk maps and residential areas. Starting from this information and using databases on real-estate values and reconstruction costs, it is possible to arrive to a damage assessment "ex ante" in case of floods useful for strategic reasoning [52]. For landslides, another Italian case is chosen. The interesting aspect of this research is that "Four types of assets are combined in the weighted assets map: (1) physical (buildings, cultural buildings and transportation networks), (2) social (population), (3) economic (land value), and (4) environmental (land cover and protected areas)" [53]. The joint evaluation of these variables on a concrete level allows us to arrive at a hypothesis that is not too far from reality.

#### **4.4 Balanced scorecard**

For monitoring and managing economic aspects of the project, we use different methods and tools that belong to the manager's toolbox, for example, cost-benefit analysis and multivariate analysis (in this work, we will focus on the former). But rather than using these raw mathematical, statistical, or financial tools alone, we integrate them in a wider framework, balanced scorecard, whose strengths are very useful for the Inte.Ri.M. project.

Balanced scorecard is a business framework used for tracking, planning, and managing an organisation's strategy. The knowledge of balanced scorecard was introduced by Robert S. Kaplan and David P. Norton in 1992 [54] and spread thanks to their book in 1996 [55]. The many criticisms the framework received are not discussed in this work.

Many benefits come from balanced scorecard. Among all, the framework allows:


Balanced scorecard is based on four "perspectives": financial, customer, internal business processes, learning, and growth. For each one, relevant topics are objectives, measures, targets, and initiatives, which must be defined, described, or computed as a preliminary step in strategy definition.

Mainly used in business, balanced scorecard is also used in the public sector, that is, government and non-profit organisations. In these cases, perspectives may be reordered differently or adapted to this different context. Specific benefits for government or non-profit organisations arise from strategy maps, a sort of add-on for balanced scorecard [56]. Strategy maps emphasise cause and effect relationships among perspectives and linked objectives but specifically take into account human capital and other variables belonging to the field of knowledge economy that

**111**

*The Internet of Things for Natural Risk Management (Inte.Ri.M.)*

project, each type of area has a distinct balanced scorecard.

**4.5 Cost-benefit analysis and cost-effectiveness analysis**

usually financial accounting does not consider. This is the most important reason why we decide to use the balanced scorecard with strategy maps. In the Inte.Ri.M.

It is in this framework that we integrate classical management tools like costbenefit analysis [57, 58], for comparing prevention costs with the damages from the disasters that can be avoided. One of the preliminary activities required to perform cost-benefit analysis is to measure all cost and benefit elements and to convert them in some unit of monetary measurement. Then, traditional indicators like net pres-

In many applications of the analysis, especially in the public sector, problems arise because elements are present that analyst or policymaker cannot or does not want to convert in a monetary measurement, that is, human lives, healthcare quality, or beauty of a landscape. For an example about environmental resources, close to our research interest, see [59]. For this reason, besides cost-benefit analysis, we also use cost-effectiveness analysis. Although typically employed in health applications, the Inte.Ri.M. project can benefit from this technique as it allows to compare alternatives based on their cost and a measure of effectiveness that is quantified but not monetised, for example, the number of lives saved instead of a monetary value of lives. Given the impossibility to compare costs expressed in monetary units and benefits expressed in another unit of measurement, cost-effectiveness analysis proceeds through the construction of cost-effectiveness indexes that allow to compare

Although the researcher's toolbox is very rich, some problems still need to be investigated. One of the most critical variables in the Inte.Ri.M. project is recovery time. Recovery time is neither a cost nor a benefit element, but it is essential to define the time window costs, and benefits are spanned over. Recovery time is difficult to estimate and is likely to increase. To make matters worse, an increase in recovery time induces changes in amount and allocation in time of monetary values. This dependence on an exogenous variable requires to reconsider the financial model and may impact considerably on results and on sensitivity analysis, one of

The chapter outlined the state of the art of Inte.Ri.M. project. The on-site activity will provide answers to the question marks determined by the specific opera-

The project will be able to generate benefits that potentially will have an impact in terms of creating value for the area. Beneficial effects are identified on two levels:

• A broad level: regarding the method of analysis development that integrates economic and technical evaluations. This output is designed to be made

• The restricted level (identified as "pilot case") that through the project will make use of hard data for risk management in the phases of prevention, intervention, and recovery is capable of enhancing the data relating to the analysis

tional conditions established by the mountain environment.

replicable in as many areas as possible.

of costs and benefits.

*DOI: http://dx.doi.org/10.5772/intechopen.81707*

ent value or internal return rate are computed.

the various alternatives.

**5. Conclusions**

the final steps of cost-benefit analysis.

*Perspectives on Risk, Assessment and Management Paradigms*

esis that is not too far from reality.

useful for the Inte.Ri.M. project.

**4.4 Balanced scorecard**

discussed in this work.

In order to monetise the value of a *human life* in disasters, we often refer to VOSL

For a "complete and detailed assessment of the exposure of 'economic activities' to natural disaster", Marin et al. study [51] refers to literature that recommends the use of different proxies depending on type of disaster. For flood, an interesting study carried out on an Italian case exploits the overlap between risk maps and residential areas. Starting from this information and using databases on real-estate values and reconstruction costs, it is possible to arrive to a damage assessment "ex ante" in case of floods useful for strategic reasoning [52]. For landslides, another Italian case is chosen. The interesting aspect of this research is that "Four types of assets are combined in the weighted assets map: (1) physical (buildings, cultural buildings and transportation networks), (2) social (population), (3) economic (land value), and (4) environmental (land cover and protected areas)" [53]. The joint evaluation of these variables on a concrete level allows us to arrive at a hypoth-

For monitoring and managing economic aspects of the project, we use different methods and tools that belong to the manager's toolbox, for example, cost-benefit analysis and multivariate analysis (in this work, we will focus on the former). But rather than using these raw mathematical, statistical, or financial tools alone, we integrate them in a wider framework, balanced scorecard, whose strengths are very

Balanced scorecard is a business framework used for tracking, planning, and managing an organisation's strategy. The knowledge of balanced scorecard was introduced by Robert S. Kaplan and David P. Norton in 1992 [54] and spread thanks to their book in 1996 [55]. The many criticisms the framework received are not

Many benefits come from balanced scorecard. Among all, the framework allows:

Balanced scorecard is based on four "perspectives": financial, customer, internal business processes, learning, and growth. For each one, relevant topics are objectives, measures, targets, and initiatives, which must be defined, described, or

Mainly used in business, balanced scorecard is also used in the public sector, that is, government and non-profit organisations. In these cases, perspectives may be reordered differently or adapted to this different context. Specific benefits for government or non-profit organisations arise from strategy maps, a sort of add-on for balanced scorecard [56]. Strategy maps emphasise cause and effect relationships among perspectives and linked objectives but specifically take into account human capital and other variables belonging to the field of knowledge economy that

• Transforming strategic objectives in operational objectives

• Measuring and checking organisation's path towards defined targets

• Improving competencies in building and implementing strategies

• Spreading knowledge of the strategy in the whole organisation

• Defining priorities of development projects and actions

computed as a preliminary step in strategy definition.

(value of statistic life) and DALY (disability-adjusted life year) [50].

**110**

usually financial accounting does not consider. This is the most important reason why we decide to use the balanced scorecard with strategy maps. In the Inte.Ri.M. project, each type of area has a distinct balanced scorecard.
