Invited perspectives: When research meets practice: challenges, opportunities, and suggestions from the implementation of the Floods Directive in the largest Italian river basin

Flood damage assessment is a non-consolidated challeng-ing river district authorities which, however, are required to produce ﬂood damage and risk maps to accomplish with the European Directive”. On the other hand, no consolidated stan-dard is for evaluations ﬂood damage assessment is still an topic scientiﬁc In a context, this paper reports challenges, opportunities, and perspectives that came to light during the current revision of ﬂood risk maps and ﬂood risk management plans (FRMPs) in the Po River District Italy; with speciﬁc reference to ﬂood damage assessment.

Authority and including several Italian universities and research centres, coordinated by Politecnico di Milano (i.e., the MOVIDA project). At present, we are in the conclusion of the project, with first results available on the webpage of the Po River District Authority (AdBPo, 2021). The objective of the project wasis to identify shared and feasible state of the art solutions for flood damage assessment in the district where, so far, damage and risk were evaluated only in qualitative terms, mostly according to expert-driven rules of thumbs (Molinari et al., 2016). To this aim the project implemented a 25 case-study, iterative approach (Figure 2). Suitable tools for the assessment of flood damage were first identified among those available in the literature and suitable for the context under investigation, or newly developed (WP_met). Such tools were then tested in 6 pilot areas, characterised by different susceptibility to be damaged in case of flood as well as data availability (WP_6); this step allowed to tune a damage assessment procedure that is flexible and usable in the different implementation scenarios, which can occur in the district. Once validated, the procedure was implemented in all the areas FRMPs ideally require consistent and comprehensive damage assessment for all items which are included in potentially heritage and environmental goodsassets. For them, we were able to assess only their amount within the potentially flooded areas, and to classify them according to some vulnerability features, linked to their susceptibility to be damaged. For the specific case of cultural heritage, given its importance in the Italian context, an ad-hoc procedure was also developed to estimate damage, even though in qualitative terms (i.e., in classes ranges from low to high damage). Indirect damage estimation was instead not feasible, although an attempt has been made to estimate consequences of roads and railways interruption. Indeed, evidence from the past (collected during the project) shows that the weight of direct damage to 55 transport infrastructures is negligible compared to the indirect one.
Paucity and low quality of georeferenced data for the evaluation of characteristics of exposed items further limited the range of damage models that could be implemented for the assessment. In fact, we dealt with scarcity of institutional databases (i.e., data are often stored in commercial repositories), legal impediments in the use of data, fragmentation of information among different databases (even for the same category of elements), and their inadequacy in supplying information required as inputs of the damage models; obsolescence of information was sometimes a problem, with data referring up to ten years ago. A meaningful example of data inadequacy for the Italian context is represented by cultural heritage, for which data are spread among several institutional databases, while a specific asset may be included in more than one database. Moreover, such databases are characterised by different structures, levels of detail, and available information; lack of metadata also hinders their interpretation and comparison.
In general, however, our experience corroborates is in line with those of the other European Member States, as can be inferred from Table 1, which summarises how flood damage is currently evaluated within the scope of the Floods Directive. In detail, Table 1 shows how different levels of analysis can currently be achieved for the various items exposed to floods, and how the MOVIDA project allowed the Po River District Authority to implement state of art flood damage 70 modelling tools in the district. In particular, the table highlights the lack of appropriate knowledge and tools for indirect damage assessment. end of the first implementation cycle, and levels of analysis achievable thanks to the tools developed in the MOVIDA project, for the various assets exposed to flood risk and kinds of damage. Characters in bold refer to ISO country codes, "MSs" stands for Member States (source: European Commission, 2016) range of damage models that could be implemented for the assessment. In fact, we dealt with scarcity of institutional databases (i.e., data are often stored in private repositories), fragmentation of information among different databases (even for the same category of elements), and their inadequacy in supplying information required as inputs of the damage models; obsolescence of information was sometimes a problem, with data referring up to ten years ago. A meaningful example for the Italian context is represented by data on cultural heritage, that are spread among several institutional 85 databases. Moreover, such databases are not complementary (duplications exist) and are characterised by different structures, levels of detail, and available information; lack of metadata also make difficult their interpretation and comparison.
Given the previous premises, the main challenge for damage assessment is the necessity to compare inhomogeneous quantities having different meanings (e.g., damage versus exposed value; direct vs indirect damage) and metrics (e.g.,

90
economic loss, physical damage, qualitative damage). Such synthesis is unavoidable if we want to assess the total impact of a flood, as an input for decision making. What our experience highlights, however, is that the close collaboration between researchers and practitioners allowed to find an equilibrium between scientific rigour and the need of technical improvement. The MOVIDA project led to the identification of feasible solutions to emerged problems and, at the same time, the transferability of scientific knowledge;

95
in this regard, the commitment of several research institutions, working together and sharing knowledge was certainly and added value. For example, thanks to such a collaborative environment we were able to develop tentative models for the estimation of indirect damage to roads and railways; we created an ad-hoc database and procedure for the assessment of damage to cultural heritage; we are presently setting up a model allowing to compare different damage-related data, by the definition of appropriate comparison criteria. The final product of the project is a comprehensive tool allowing for of flood mortality, indirect damage to people, direct and indirect damage to economic activities, damage to cultural and environmental goodsassets, and damage to infrastructures, also with specific reference to coastal areas (WP_svi in Figure   2)..
What clearly emerged from our experience is that flood damage (and, more in general, risk) assessment is not a solely technical problem; social and economic aspects are key elements, calling for a multidisciplinary and a participative approach. Local stakeholders must be especially involved in the final synthesis of the damage evaluations (WP_tav in Figure 2): as previously discussed exposed assets are evaluated through different metrics and suffer from different type of consequences as a result of a flood; the total damage must, therefore, reflect the perception of such values by those who will make use of the assessment in decision-making.
The main conclusion that we can infer from the development of the MOVIDA project is that implementing available and new-coming damage models in real practice is the most appropriate way towards the standardisation of damage 115 assessment tools. Indeed, differently from other disciplines, flood damage models cannot be validated by laboratory tests.
Their quality, validity and transferability must be evaluated on the field, and strongly depend on the objective for which the model is implemented, as performing damage assessment for long-term planning purposes may have different requirements than for insurance, or emergency management related reasons. In fact, a model can be very useful for one objective and not for others. To clarify whether or notwhether a damage model is useful, scientists, practitioners and 120 stakeholders must confront each other, overcoming shared practices. The MOVIDA project represented a good opportunity in this direction. We wish that such a collaborative way of working will be adopted not only in other districts or river basins, but at the European community level. Indeed, in light of the harmonisation required by the European Commission as regards the implementation of the Floods Directive among member states, a comparison between scientists, practitioners and stakeholders at the European level would be suitable, in order to converge on objectives and 125 methods and, in turn, on homogenous requirements of input data at the European level (on which improved datasets can be designed), in a top-down approach; . Table 1 shows instead a very fragmented reality as regards flood damage assessment and mapping at the European scale, where the level of analysis achievable for the different exposed assets strongly depends on the availability of national/local tools and required input data. As occurred during the MOVIDA project, the new research challenges and directions will consequently emerge. We claim that such an approach would be 130 beneficial not only for damage assessment related problems but also for challenges linked to other aspects of flood risk management (like climate change), or to the management of other risks. Our wish is the creation of real opportunities to work in this direction, as the definition of a European platform, a COST action or, more ambitiously, an inclusive, big (research) project supported by the European Commission.