Vulnerability curves vs . vulnerability indicators : application of an indicator-based methodology for debris-flow hazards

The assessment of the physical vulnerability of elements at risk as part of the risk analysis is an essential aspect for the development of strategies and structural measures for risk reduction. Understanding, analysing and, if possible, quantifying physical vulnerability is a prerequisite for designing strategies and adopting tools for its reduction. The most common methods for assessing physical vulnerability are vulnerability matrices, vulnerability curves and vulnerability indicators; however, in most of the cases, these methods are used in a conflicting way rather than in combination. The article focuses on two of these methods: vulnerability curves and vulnerability indicators. Vulnerability curves express physical vulnerability as a function of the intensity of the process and the degree of loss, considering, in individual cases only, some structural characteristics of the affected buildings. However, a considerable amount of studies argue that vulnerability assessment should focus on the identification of these variables that influence the vulnerability of an element at risk (vulnerability indicators). In this study, an indicator-based methodology (IBM) for mountain hazards including debris flow (Kappes et al., 2012) is applied to a case study for debris flows in South Tyrol, where in the past a vulnerability curve has been developed. The relatively “new” indicator-based method is being scrutinised and recommendations for its improvement are outlined. The comparison of the two methodological approaches and their results is challenging since both methodological approaches deal with vulnerability in a different way. However, it is still possible to highlight their weaknesses and strengths, show clearly that both methodologies are necessary for the assessment of physical vulnerability and provide a preliminary “holistic methodological framework” for physical vulnerability assessment showing how the two approaches may be used in combination in the future.

results, however, it still bears many uncertainties. The same authors also stress that challenges are related to the selection of indicators, their standardisation, the availability of required data, their weighting and the methods of aggregation for the development of a vulnerability index, concluding that empirical investigation (e.g. vulnerability curves) may lead to better results than the use of vulnerability indicators. However, vulnerability indicators are very often used for the assessment of social vulnerability as for example in the Social Vulnerability Index of Cutter et al.(2003). In the case of physical vulnerability, 5 although vulnerability curves are more popular than vulnerability indicators, studies using indicators for assessing physical vulnerability are on the rise. Often, some of these studies focus on the development of inventories of elements at risk and their characteristics (Papathoma-Köhle et al., 2007) or they make an additional step towards exposure assessment (Fuchs et al., 2015). Barroca et al. (2006) developed a vulnerability analysis tool for floods based on a system of indicators describing the elements at risks, their spatial relation, the prevention, emergency and reconstruction systems. Barroca et al. (2006) suggest 10 that the vulnerability assessment tool is simple and flexible and can be used by different users without requiring expert knowledge for its use. On the other hand, Müller et al. (2011) assessed urban vulnerability towards flood using indicators for physical and social vulnerability. Concerning physical vulnerability, four indicators were selected (construction material, building position, proportion of green spaces and local structural protection) and ranked based on expert judgement. Most of the indicator-based methodologies found in the literature are applied at local scale. The study of (Balica et al., 2009) is 15 noteworthy because the selected indicators used to develop indices for flood vulnerability are available for three scales (riverbasin, sub-catchment and urban area). Balica et al. (2009) suggest that the indicator-based methodology is powerful because it supports the decision makers in prioritization and it supports transparency. Apart from floods, physical vulnerability has been investigated with means of indicators also for other hazard types such as landslides. Silva and Pereira (2014) assess physical vulnerability of buildings to landslides based on the building resistance and the landslide magnitude. The building's 20 resistance is determined by a number of indicators including construction technique and material, number of floors, floor and roof structure and conservation status. Last but not least Kappes et al (2012) developed a new indicator-based methodology for multi-hazard in mountain areas which is going to be presented in detail and applied in the present study. In general, in most of the studies found in the literature using indicators, the indicators are weighted empirically and no validation of the indicators selection and weighting has been implemented using the damage pattern and loss from a real event. 25 2.2 The PTVA method and its evolution One of the first attempts to use indicators for the assessment of physical vulnerability was made by Papathoma and Dominey-Howes (2003) and Papathoma-Köhle et al (2003). Specifically, they developed a methodology for assessing the physical 30 vulnerability of buildings to tsunami at coastal areas in Greece using indicators. The main concept of the Papathoma Tsunami Vulnerability Assessment Model (PTVA) was the combination of an inundation scenario with "attributes relating to the design, condition and surroundings of the building" (Tarbotton et al., 2012). The methodology was based on the fact that two buildings located exactly at the same place despite experiencing the same process intensity do not always suffer the same loss. The Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-76, 2016 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. reason for this is the variety of building characteristics concerning the building itself and its surroundings. A number of buildings characteristics related to the damage pattern following a tsunami event were selected (indicators) and a GIS database with the buildings and their attributes was created. The indicators were weighted using expert judgement and a vulnerability index was given to each building. The result was a series of maps for a number of coastal segments in Greece showing the spatial pattern of the relative physical vulnerability of the buildings. In this way, authorities and emergency services could 5 focus their limited resources in specific buildings rather than to the whole potentially inundated area. The method was later validated (Dominey-Howes and Papathoma 2007) using data from Maldives following the 2004 Indian Ocean tsunami. The validation showed that the selected vulnerability indicators correlate well with the severity of the damage, however, recommendations for improvement led to an improved version of the method: PTVA-2. PTVA-2 was used in USA (Dominey-Howes et al., 2010) for the estimation of the probable maximum loss from a Cascadia tsunami in Oregon (USA). The main 10 difference to PTVA-1 was the inclusion of the water depth above ground as an attribute in the calculation of the overall vulnerability. In this way, the method became more intensity dependant than before. PTVA-3 was developed by Dall' Osso et al (2009a) and was tested in Australia (Dall'Osso et al., 2009b) and in Italy (Dall'Osso et al., 2010). PTVA-3 made a step towards and more reliable weighting of the vulnerability indicators by using Analytical Hierarchy Process (AHP) rather than expert judgement to weight the attributes. 15 The first attempt to use vulnerability indicators for mountain hazards was made by (Papathoma-Köhle et al., 2007) with the development of an element at risk database containing indicators regarding the physical vulnerability of buildings to landslides.
However, due to the lack of data regarding the hazardous process itself and the limited availability of data regarding the building characteristics the study constitutes only a good basis for further research. The next attempt was made by Kappes et 20 al. (2012) introducing a methodology for vulnerability assessment using vulnerability indicators, e.g. characteristics of the building that are responsible for its susceptibility to damage and loss due to mountain hazards, based on the PTVA model. The methodology is presented in the following chapter. It is clear that as far as mountain hazards are concerned, there is definitely a need to improve and modify indicator-based approaches in order to be used for vulnerability assessment and as basis for risk reduction strategies. 25 The two methodological concepts have been always used in conflict rather than in combination from scientists and practitioners. A first effort to combine them has been made by (Papathoma-Köhle et al., 2015). Papathoma-Köhle et al. (2015) developed a tool which uses the vulnerability curve presented in the following chapters as a core to implement three functions: updating and improvement of the curve with data form new events, damage and loss assessment for future events, damage 30 documentation of new events. However, the tool has the possibility to include information regarding building characteristics.
This provides the opportunity to investigate the correlation of damage patterns and the building characteristics in the future.

The indicator based methodology: a PTVA for debris flow
The concept of the indicator-based methodology for mountain hazards (debris flow, landslides and floods) is based on the assignment of weights to a number of building characteristics resulting to a Relative Vulnerability Index (RVI) per building ( Figure 1). The RVI is calculated per building by using the following equation (1): Where w represents the different weights, I the indicators and s the sores of the indicators as shown in Figure 1. A selection of vulnerability indicators of buildings for debris flow events and the weighting according to different users are shown in Figure   2. In more detail, the vulnerability indicators are related to building characteristics (material, condition, number of floors, building surroundings) and location (row towards torrent and row towards the slope). The indicators are weighted and aggregated according to equation (1) to a Relative Vulnerability Index (RVI) which is attributed to each building. 10 However, the weighting is not static since the needs of the end-users may vary. Therefore, each user of the method should be able to set their own priorities and change the weighting according to them. For example, emergency services need to know the physical vulnerability of the buildings in order to locate concentrated number of casualties and potential victims following an event. Therefore, buildings not offering vertical evacuation opportunities (one floor buildings) are of higher importance. 15 On the other hand, if the aim of the study is the development of vulnerability reduction efforts during the preparedness phase, building material and condition or the existence of building surroundings are of the highest importance. In this way, several RVIs may be calculated for each building: for example, an RVI RP, which can be used for emergency planning, or an RVI VR which can be used as a base for vulnerability reduction strategies (e.g. reinforcement of buildings). Kappes et al. (2012) suggest that the main advantage of the method is the flexibility in weighting. Moreover, they consider the fact that the method is not hazard intensity specific to be an advantage, because the assessment may be carried out in absence 5 of information regarding the process characteristics. The approach is based on a "relative" vulnerability index which more or less highlights the buildings that are more vulnerable than the others. Kappes et al (2012) based on the limitations of the method outline the necessary actions that have to be taken in the future in order to improve the method. They recommend that damages of events have to be recorded in a more detailed way in order to comprehend the role of each indicator and their importance in determining the vulnerability of the building. Since the method requires a large amount of detailed data 10 alternative data collection methods may be introduced, e.g. questionnaires, google street view etc. Additional data, such data regarding the open spaces, the accumulation of movable objects or even additional elements at risk such as agricultural spaces and industry, will make the method more integrated. Furthermore, the database could be enriched with socio-economic data at building level. Last but not least, Kappes et al. (2012) suggest that an interesting development of the method would be the validation of the indicators weighting based on damage records of past events.

The case study area 5
The indicator based methodology has to be tested in an area that not only has a long record of debris flow events but also detailed documented damages on buildings that may reveal information regarding the intensity of the event itself but also the damage pattern on the built environment. For this reason, the methodology will be validated in Martell (South Tyrol, Italy).
Communities in South Tyrol (Italy) often suffered material damages due to geomorphological hazards such as landslides, flash floods and debris flows.  Additionally, a reservoir dam was constructed in 1956, which served mainly as an electrical power source and protected the village from unexpected excessive flooding.
On 24 August 1987, following some days of continuous and heavy precipitation, the river Plima transported a significantly higher amount of water than usual. Debris flows were initiated in tributary streams along the valley. In the evening, the 5 inhabitants were successfully evacuated, as the water level continued to rise. A couple of hours after the evacuation, a large debris flow went through the valley, causing devastation. The specific debris flow event cannot be considered as entirely natural, as it was directly connected to the mismanagement of the reservoir dam which failed to regulate the water flow into the valley (Pfitscher, 1996). The actual outflow has been estimated to be three times the usual discharge (300 to 350m3/s). The debris flow reached the village of Gand, overflowed the river bed and found its way through the settlements. Not only buildings 10 were destroyed, but significant damages were recorded in agricultural areas were flooded, infrastructure and the industrial zone in Vinschgau (Pfitscher, 1996). Fortunately, due to the early warning and evacuation no casualties were recorded.
Although the total damage summed up to 45 to 50 billion lire (approximately 23.2 to 25.8 million EUR), private households suffered damages of slightly less than 8 million EUR. As far as the infrastructure, local light industry, forestry, agriculture, tourism and emergency response are concerned, damages up to 4 million EUR were recorded. Furthermore, 5.6 million EUR had to be spent for the recovery of the regional road and the telecommunications network as well as for torrent control measures. At the time only the direct costs of the event were 20 assessed (Pfitscher, 1996). The event was documented mainly through photographic material ( Figure 4) that enabled later on the assessment of the intensity of the process as well as the damage pattern on individual buildings.

The development of a vulnerability curve for the study area
A vulnerability curve was developed based on empirical damage data of buildings in Martel, South Tyrol, Italy, that were damaged during the 1987 event ( Figure 5). The event was not thoroughly documented, however, following the event the buildings, as well as parts of the affected villages, were photographed. The photographs give a very good overview of the damage pattern, however, information regarding the costs of the damages per building, or complete description of the damage 5 were not available. In more detail, photographic documentation of 51 buildings out of the 69 buildings that were damaged or completely destroyed during the event was used (Pfitscher, 1996), because only for this amount of buildings adequate photographic documentation was available. Based on these photos, the height of the debris deposits per building could be estimated and the monetary damage was calculated. For example, in case the damage is limited to the exterior of the building, the works will include cleaning, plastering and painting of the external walls. In case the debris has entered the building through 10 openings or wall breaks, there will be additional renovation works such as removal of debris and cleaning of rooms, restoration of interior walls and floor, testing of the electricity and heating network etc. The vulnerability curve clearly shows that the higher the intensity of the process the greater the damage that an element at risk suffers. The curve indicates that if the intensity exceeds 1.5m, the degree of loss increases considerably. This may be explained by the fact that debris flow of this height may easily enter the building from windows and doors and cause additional damage in the interior. However, at this point, it is important to emphasise that only the structural damage was considered in the 5 calculation of the monetary loss and not the content of the buildings. values were recorded, probably because roads acted as corridors that enabled the debris flow to enter the settlement area.
Moreover, in Figure 6 is also obvious that in Ennewasser (north of the map) the intensities were significantly lower due to the protection offered by the forest on the east side of the settlement. In Figure 7 the spatial distribution of the degree of loss follows in most cases the pattern of the intensity distribution. As Table 1 clearly shows the majority of the buildings experienced rather low intensity of debris flow (less than 1m debris height). Only 9 buildings (17,5%) experienced high 20 intensity of debris flow (more than 2m debris height).

The application of the indicator-based methodology in the study area
The indicator-based methodology was applied at each building that experienced damage during the event of 1987 in Gand and Ennewasser (Municipality of Martell). Based on photographic documentation provided form the municipality of Martell a GIS a database was developed containing the vulnerability indicators of each building. Information regarding the material, the building condition and the intensity could be collected from photos, whereas, information regarding the building surroundings 5 and the building's location in relation to the neighbouring buildings could be acquired from an ortho-photo of the area. An example of the calculation of physical vulnerability using indicators is shown in Figure 8.

Figure 9
The spatial distribution of physical vulnerability using indicators for emergency planners Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-76, 2016 Manuscript under review for journal Nat. Hazards Earth Syst. Sci.  Figure   7) are often the ones with a high RVI especially in the case of RVI VR . However, the visual comparison of the maps shows that some buildings that experienced a very high degree of loss have been assigned with a low vulnerability index and vice versa.
For example, the buildings in Ennewasser, that have experienced low intensities and for this reason they also showed a relative 15 lower degree of loss, were classified as highly vulnerable by the indicator based method. This observation highlights an important aspect of the indicator-based method which is the fact that the indicator based methodology is not hazard intensity specific in contrast to the vulnerability curve method that includes information regarding the intensity. On the other hand, there are buildings in Gand that although they have experienced a high degree of loss they are assigned with a low RVI. The high degree of loss it is expected since according to Figure 6 the same building also experienced high intensity. It is, therefore, clear 20 the results show inconsistencies because the intensity is not considered by the indicator-based methodology. However, there are building characteristics that are directly connected to the intensity that a building experiences e.g. surrounding vegetation or protection from other buildings, proximity to the road network etc.
By visualising the spatial pattern of the degree of loss and the observed intensity, additional valuable information regarding 25 the importance of vulnerability indicators may become available and, in this way, the indicator-based methodology may be modified, extended and substantially improved. For example, the low intensities of recorded on buildings in Ennewasser may be related to the presence of surrounding vegetation which highlights the importance of the relevant vulnerability indicator.
This leads to the conclusion that both methods, although based on a different concept they shed light in two different aspects of the physical vulnerability, namely the characteristics of the buildings and the expected degree of loss under a given intensity 30 respectively and, therefore, they should be used in combination rather than in conflict.
However, additionally to the visual comparison of the results, a closer look to the results for each building in Table 2 reveals even more about the advantages and disadvantage of both methods.  This is to be expected since coincidently there are no buildings in the area with extreme scores (e.g. material=wood or condition=ruin).
2. There are no buildings with RVI=0. That would mean that a building is not vulnerable. However, the assignment of zero scores that could lead in such a result is not possible. This is understood since no building can be characterised 5 by zero vulnerability since it is located within the area affected by the debris flow. This is also supported by the vulnerability curve which is based on a real event and shows that no building had zero degree of loss.
3. As far as the comparison between the two methodologies is concerned, some buildings although they have experienced low degree of loss and they are assigned with a high RVI. This can be explained by the fact that the indicator based methodology is not intensity specific. The high RVI means that the specific building could experience 10 high degree of loss due to its characteristics. On the other hand, the vulnerability curve shows that the specific building did not experience high degree of loss during the specific event for reasons that maybe connected to characteristics of the process itself (e.g. G48). This is even more obvious in the case of buildings G53 and G56. The two buildings have been assigned with similar RVIs. They have, however, experienced degree of loss 1 and 0,27 respectively. This can be explained by the significant difference in the process intensity (1,7m and 3m respectively). 4. However, the opposite observation is also clear: some buildings have experienced high intensity which resulted in a low degree of loss (e.g. G53, G64, G66, G224, G226). Obviously the fact that buildings although they experience the same intensity of the process they experience different degree of loss (some are completely destroyed and some others have less than 50% of loss may be explained by differences in their characteristics. However, another explanation may be characteristics of the process that are not considered by the vulnerability curve such as velocity of the flow, 20 direction of impact, viscosity, duration, size of debris etc. The results of the indicator methodology depend on the set of indicators used for the calculation of the RVI. The indicatorbased methodology is based on a set of indicators that were selected based on expert judgement, damage reports and photographic documentation of past events. However, Birkmann (2006) based on existing set of indicators (EEA, NZ Statistics 25 etc.) provides a list with quality criteria that vulnerability indicators have to fulfil. In Table 3 the set of indicators of the present study is tested towards some of these criteria.

Measurable
The indicators used are not always easily measurable. The difference between a building of "good" or "medium" condition is not clear and not measurable in quantitative terms. The scores for each building may be dependent on the judgement of the data collector and may not always be objective. However, improved data collection techniques (e.g. detailed standardised questionnaires etc.) may improve the measurability of the indicators.

Relevant
The indicators have been based on reports and documentation of past events and for this reason are relevant to the assessment. They have also been chosen according to the needs of the end-users although the latter could be more involved in the selection process in the future. The weighting, however, is done directly by the end-users.

Policy-relevant
Although not demonstrated in the specific article, the indicators may be policy relevant. The vulnerability indicators may give decisions makers an overview of damage potential for future events.
Moreover, they may be used for emergency planning and they may guide local structural protection measures.

Measure important
The indicators are connected to key elements (e.g. reaction of the structure to the impact of debris flow) and are not attempting to indicate all aspects (e.g. vulnerability to other hazard types).

Analytically and statistically sound
Although the indicators are analytically sound in the sense that they may depict the actual situation with accuracy, the links between natural process and degree of loss as well as the reaction of a structure to the natural process according to its characteristics are not fully understood and therefore, further research is required.

Easy to interpret
The indicators used in this study are easy to interpret. No expert stuff is required for their collection.
Although this is an advantage of the method, the judgement of the collector may influence the result significantly.

Sensitivity
Although the indicators are specific to the phenomenon of debris flow, they are not sensitive in changes of this phenomenon such as its intensity. However, they are sensitive to changes in the structure of the building which means that they are able to express changes in the physical vulnerability should a building is reinforced.
Validity/accuracy The indicators have the capacity to express the physical vulnerability of buildings in most of the cases and this may be also confirmed by the results of the vulnerability curve. However, this is not always the case. The cases where high vulnerability has been assigned for buildings that have experienced low degree of loss have to be investigated and based on the conclusions of this investigation the methodology has to be improved.

Reproducible
Theoretically, the indicators set could be reproducible for another area facing a threat of debris flows.
However, significant differences in the architecture and building standards of buildings should be considered.
Based on available data The indicators are not always based on available data. Some information could be available from the municipality. However, the majority of the required information may be collected through field work or interpretation of ortho-photos.

Data comparability
The indicators may be compared to similar ones in other areas but also to past conditions of the same area of future hypothetical ones.

Cost effective
The indicators are cost effective. The assessment of the vulnerability for debris flow usually involves a limited amount of buildings. Although fieldwork is necessary there are ways to avoid it by sending questionnaires to the building owners or interpreting ortho-photos for rapid data collection. Table 2 shows that although the set of indicators fulfils many criteria, there is still room for improvement of the indicator set itself, the description of the attributes for each indicator and the collection of the required data.

Benefits, limitations and future development of the indicator-based method 5
In contrast to the vulnerability curves, indicator-based methodologies are not established and applied by practitioners in the same dimension. For this reason, the comparison between the two concepts is essential in order to highlight the benefits of the indicator-based methodologies, point out its weaknesses and make a list of recommendations not only for the improvement of the indicator-based concept but for the improvement of the assessment of physical vulnerability in the future. 10 The comparison of the two methods highlighted the following benefits of using indicators for assessing physical vulnerability of buildings to torrential hazards: 1. In contrast to the development of vulnerability curves, the indicator based methodology prompt the user to develop an inventory of the elements at risk and a database of building characteristics. This may enable the development of strategies for the reduction of vulnerability at a very local level, as well as the development of local structural 15 protection measures. In this way the focus and the resources will be concentrated on limited amount of buildings.
2. The indicator-based methodology does not use empirical loss data but it rather indicates the relative vulnerability of individual buildings. This means that this type of methodologies can be applied to areas with no recorded history of events making their use easier in absence of damage data. easy to update databases future changes in the spatial pattern of the built environment, socio economic and land use changes may be considered for future scenarios.
6. The use of GIS enables the visualisation of the spatial pattern of the physical vulnerability and also individual characteristics. In this way vulnerability maps may be used as a basis for e.g. emergency planning. On the contrary, vulnerability curves do not have a spatial component. 5 7. The transferability of methods in the field of risk research is often due to differences in the nature of elements at risk, environmental conditions and processes impossible. However, since the specific method is not intensity and process dependant, it is easier to transfer to other areas with similar problem provided that necessary modifications will be made (e.g. additional building characteristics due to local architecture).
8. Indicator-based methods encourage the involvement of local communities and individual building owners in data 10 collection and vulnerability reduction.
However, indicator-based methodologies for physical vulnerability assessment are in their infancy allowing a large space for improvement. This improvement requires the identification of the main drawbacks of the methodology: 15 1. Intensity relevance: the indicator methodology assigns a relevant vulnerability index to each building which more or less shows which building is more vulnerable than the other in a worse-case scenario without indicating a specific intensity of the event. This is a major drawback which is obvious from Table 2: buildings with high RVI experienced a very low degree of loss. A possible development of the method could be similar to the one of PTVA-2 which included the water height (in this case the deposit height) as an indicator in the vulnerability assessment, 20 based on a specific scenario. Table 2 is obvious that there some inconsistencies between the two methods for specific buildings. The reason may be the lack of completeness of the set of indicators. For example, characteristics that may affect the vulnerability of the building significantly, such as the existence of openings on the slope side, their size and quality, as well as the existence of a basement are not considered. 25 3. Completeness of data sets and costs of data collection: The data set required for the implementation of the methodology is detailed and has to be collected at local level.

Description of scores:
The scores for the various indicators are often described in a trivial way. E.g. "good" or "medium" building condition leading to a large dependence on the judgement of the data collector to decide the score that will be assigned to a building. 6. "Relativeness" of the vulnerability indicator: The RVI expresses the relative vulnerability of buildings in an area. That means that the RVI points out which building is more vulnerable than the other without having the capacity to translate this vulnerability into a quantitative value. This may be considered a disadvantage for practitioners because the vulnerability map made this way may only indicate vulnerable buildings in an abstract way and vulnerability maps from different areas may not be compared. 5 7. Uncertainties: Both methods bear a number of uncertainties that should be analysed and quantified. As far as the end-users are concern, decision makers are in need of vulnerability assessment methods that they can use for risk analysis but they also need to know the uncertainties that are associated with the vulnerability values. This is important because low estimated risk with high uncertainties may cause higher losses than medium estimated risks d. Credibility of existing data: Some buildings although they were not severely damaged got full compensations to be rebuilt due to relocation. 20 As far as the indicator-based method is concerned, uncertainties are related to the following: a. The subjectivity of the data collector (e.g. what is a "good" and what is a "medium building condition?) b. The subjectivity of the end-user concerning the weighting and the classification method.
Efforts to analyse and quantify uncertainties concerning the assessment of physical vulnerability for debris flows have been made in the past ((Totschnig andFuchs, 2013),(Eidsvig et al., 2014)), however, most of them concern 25 vulnerability curves.
Based on the comparison of the two methods and the outline of the advantages and disadvantages of it as far as the assessment of physical vulnerability is concerned a number of recommendations for improvement may be made: 1. Reduction of data collection effort and time through the use of modern technologies (GIS and remote sensing): 30 Information regarding the surroundings or the building row is easier to be collected by using GIS maps and/or remote sensing where needed. For other type of information such as openings, existence of basement and building material and condition an additional field survey may be necessary. Alternative data collection methods (e.g. distribution of standardised questionnaires to the building owners) may also be introduced.
Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-76, 2016 Manuscript under review for journal Nat. 2. Improvement of post-disaster documentation methods: Detailed post disaster documentation including photographic material may provide valuable information regarding the interaction between natural processes and elements at risk but it may also give information regarding the variation of the intensity of a process within a given area. An improved method for damage documentation is recommended by Papathoma-Köhle et al (2015).
3. Reconsider the score description: some scores would be more reliable if they would be less dependent on expert 5 judgment (for example building condition). A solution would be to reconsider their descriptions and come up with tangible scores (e.g. age of the building, or detailed description of building condition).

Additional indicators:
The list of indicators used for the assessment of physical vulnerability of buildings to debris flow is not exhausted. Information regarding the presence of a basement, the number location, size and quality of openings is also missing. Furthermore, information regarding the local structural protection of buildings (e.g. 10 elevation, splitting wedges, deflection walls etc.) should also be included.

5.
Physical resilience: Some of the indicators mentioned above may also contribute not only to the assessment of the physical vulnerability of the building but also to the assessment of its physical resilience. According to Haigh (2010, p.4) resilient buildings should "enable society to continue functioning when subject to a hazard". In this respect, if the interior of the building has been invaded by debris or if the heating and electricity network are located 15 in the basement and have been heavily damaged by intruding water and material the building will need more time to be re-inhabited by its occupants. Therefore, indicators regarding the physical resilience of buildings may include location of vital equipment, existence of openings, local protection measures, basement, primary or secondary residence of the inhabitants).
6. The interaction of structures with the natural process has not been thoroughly investigated. Gems et al. (2016)  A database which includes indicators related to more than one hazard type would be a good start. 30 Improving indicator-based methodologies is an important step towards disaster risk reduction, however, the advantages of the vulnerability curves are also indisputable, a comparison of the two methods may shed light to their advantages and drawbacks and may also inform the practitioners on their available methodological choices. However, as it is suggested in the following paragraphs both methodological approaches are needed.

Vulnerability curves vs. vulnerability indicators
The above comparison of the two methodological concepts clearly shows that decision-makers and other potential end-users are actually in need of both methods during the different phases of the disaster management cycle for different reasons. In 5 short, decision makers, authorities and disaster managers need a quantitative representation of physical vulnerability based on empirical data (curves) as well as, relative vulnerability values for individual elements that are directly connected to their characteristics (indicators) and may support prioritisation of resources and small local interventions for vulnerability reduction.
The fact that both approaches are significant stresses the need for a "holistic physical vulnerability framework for practitioners" that will enable practitioners to choose the relevant method according to their aim and the available resources. An effort to 10 place the various methodological concepts (including in this case also vulnerability matrices) within the phases of the disaster cycle has been made by (Papathoma-Köhle and Ciurean, 2014).  Figure 11 shows the importance of vulnerability assessment in every phase of the disaster cycle. However, for each phase a different method is relevant. For example, during the mitigation phase vulnerability curves may be used for the loss estimation of future events and the design of mitigation measures. The results in this case may be used for public awareness and education but also for recommendations for reinforcement. On the other hand, vulnerability maps based on the vulnerability indicators 5 may guide the emergency and evacuation planning process during the preparedness phase. Additionally, during the response phase, vulnerability maps are essential to indicate the buildings that are more possible to have been damaged, whereas the recovery phase is the ideal phase for validating the weighting of indicators. Moreover, during the recovery phase the results of the vulnerability assessment may be used as guidance for relocation of buildings. Furthermore, it is essential that methods and tools for analysing and quantifying uncertainties should be also included in a "holistic physical vulnerability framework for 10 practitioners". The scheme in Figure 11 could form the basis of a framework for practitioners, however, further work is needed on this respect.

Conclusions
Vulnerability assessment constitutes a large part of risk analysis and its reduction has a direct effect on the consequences of 15 natural disasters on communities, buildings and infrastructure. The variety of methods available are used until now mostly in isolation and there is often a debate about the advantages and disadvantages of different methods. In this chapter, through the comparison of two of the most common vulnerability assessment approaches (vulnerability indicators and vulnerability curves) a new methodology is applied and validated. The vulnerability curves although widely used by practitioners they reveal the need to include the characteristics of the buildings (indicators) in the assessment of physical vulnerability. The indicator-based 20 method presented and applied in the present article shows that although assessing physical vulnerability using indicators may give reliable results, there is significant work that has to be done in order to improve indicator-based methodologies for physical vulnerability assessment. Moreover, the article emphasizes the need for a "holistic framework for physical vulnerability assessment for practitioners" that will offer the opportunity to the end-users to use both methodologies in a complementary way, quantify uncertainties and consider change, not only as far as the natural process is concerned but also socio-economic 25 and land use changes that will have a direct influence on the consequences of natural disasters.

ACKNOWLEDGMENTS
This project received funding form the Austrian Science Fund (FWF): P 27400. Data collection was partly supported by the EU project MOVE (Methods for the Improvement of Vulnerability Assessment in Europe) (contract number 211590). The 30 author would like to thank the Local Authorities of Bolzano and the Municipality of Martell for the provision of data and photographic material. Furthermore, Thomas Thaler and Sven Fuchs are thanked for the discussions on earlier drafts of the present article. Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-76, 2016 Manuscript under review for journal Nat. Hazards Earth Syst. Sci.  Natural Hazards, 64, 1969-1975, 2012 Fuchs, S., Heiss, K., & Hübl, J.: Towards an empirical vulnerability function for use in debris flow risk assessment. Natural 30 hazard and Earth System Sciences, 7, 495-506, 2007. Fuchs, S., Keiler, M., & Zischg, A.: A spatiotemporal multi-hazard exposure assessment based on property data. Natural Hazard and Earth System Sciences, 15, 2127-2142, 2015. Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-76, 2016 Manuscript under review for journal Nat. Hazards Earth Syst. Sci.