EUROPEAN COMMISSION

Brussels, 15.12.2021

SWD(2021) 971 final

COMMISSION STAFF WORKING DOCUMENT

European Overview- 2nd Preliminary Flood Risk Assessments

Accompanying the document

REPORT FROM THE COMMISSION TO THE EUROPEAN PARLIAMENT AND THE COUNCIL

on the implementation of the Water Framework Directive (2000/60/EC), the Environmental Quality Standards Directive (2008/105/EC amended by Directive 2013/39/EU) and the Floods Directive (2007/60/EC)









Implementation of planned Programmes of Measures





New Priority Substances





Preliminary Flood Risk Assessments and Areas of Potential Significant Flood Risk

{COM(2021) 970 final} - {SWD(2021) 970 final}

Contents

1.Executive Summary4

2.Introduction and background5

3.Overview of timeliness and completeness of the information reported6

3.1.Timeliness of the reported information7

3.2.Completeness of the reported information9

4.UoMs and Competent Authorities11

5.Implementation of Article 412

5.1.Introduction to Article 4.2 of the Floods Directive12

5.2.Article 4.2(a) – Maps with topography and land use17

5.3.Article 4.2(b) – Assessment of past floods with significant impacts19

5.3.1.General information on past floods20

5.3.2.Sources, mechanisms and characteristics of past floods22

5.3.3.Extent and frequency or recurrence of past floods24

5.3.4.Quantitative data on impacts of past floods27

5.3.5.Methodologies and criteria used for the assessment of the significance of the impacts of past floods32

5.4.Article 4.2(c) – Assessment of past floods without known significant impacts that may have significant impacts if repeated in the future40

5.4.1.Methodologies and criteria used for the assessment of significant past floods without known impacts40

5.5.Article 4.2(d) - Assessment of future floods with potential adverse consequences regardless of significance45

5.5.1.Sources, mechanisms and characteristics of future floods45

5.5.2.Methodologies and criteria used for the assessment of the significance of the consequences of future floods46

5.5.3.Potential adverse consequences of future floods52

5.5.4.Issues considered in the assessment of adverse consequences of future floods54

5.6.PFRA aspects of special interest59

5.6.1.Pluvial floods (flash floods) in urban settings59

5.6.2.Inclusion/exclusion of floods from sewerage systems59

5.7.Changes in Article 4’s assessments since the previous cycle63

5.7.1.Article 4.2(b) – developments since the previous cycle63

5.7.2.Article 4.2(d) -developments since the previous cycle64

5.7.3.Changes to the methodologies of Articles 4.2(b), 2(c) and 4.2(c) since the previous cycle65

5.7.4.Distinction between Articles 4.2(b) and 4.2(c) or 4.2(c) and 4.2(d)67

6.Implementation of Article 5 - APSFRs68

6.1.Methodologies and criteria used for the selection of APSFRs68

6.2.APSFR selection– developments since the previous cycle71

6.3.Number of APSFRS and changes to APSFRs since the previous reporting74

6.4.Sources, mechanisms and characteristics of floods in ASPFRs76

6.5.Adverse consequences associated with APSFRs82

7.Consideration of climate change85

7.1.Evidence of consideration of climate change by the Member States85

7.2.Development of the consideration of climate change in future PFRAs87

List of Acronyms

List of Member States’ codes

1.Executive Summary

This document is an European Union overview of the Member States’ updated preliminary flood risk assessments, and identification of areas of potential significant flood risk, according to Articles 4, 5, 14 and 15 of the Floods Directive. These updates were to be reported to the European Commission by March 2019. The document brings together, and discusses, the findings of a review conducted by the Commission that examined each Member State’s update individually. The findings for each Member State are published in separate documents. At the time of publication, all Member States have reported information on the implementation of this part of the Floods Directive through the European Environment Agency’s Water Information System for Europe. It was not possible to include the update of one Member State, since it reported very late to be included.

The present EU overview aims at strengthening flood risk management in the EU The Commission also collects information to create an EU-wide picture to inform the public on certain aspects of policy. The present document may take therefore a broader perspective than the Floods Directive; the text of the latter being the only benchmark against which a Member States’ compliance should be judged.

None of the Member States have made any notable changes to their administrative arrangements. Reports detailing the updated preliminary flood risk assessment have been prepared by all Member States covering all river basins. Nearly all Member States published their preliminary flood risk assessments online.

Overall, compared to the Member States’ first preliminary flood risk assessments, half have improved data collection and/or methodologies to carry out preliminary flood risk assessments. In the previous Commission review, no distinction was drawn between the methodologies for the application of the various sub-articles under Article 4. This has now become clearer, however, there is still room for improvement. Therefore, Member States should consider providing clearer information on how Article 4 has been applied in the next update of their preliminary flood risk assessment. To aid this process, a flowchart detailing the possible steps involved is included in this document.

Although the discourse around floods in urban areas and sea level rise has intensified, it is still river floods that are most frequently registered as a source of significant flooding in the EU. The most common mechanism of floods happening was natural exceedance (of e.g. the confines of a river’s banks or embankments) and the most common characteristic was flash flooding, i.e. flooding that materialises rather quickly.

All Member States provided at least some information on how past floods have been assessed and the criteria used for defining significance. In some Member States detailed information on how the criteria and methodologies have been applied are lacking, but in others the methodology is clear and detailed. In fact, two thirds of Member States presented strong evidence of a clear methodology for the assessment of past floods. In addition, “expert judgement” has been relied upon to a lesser extent, mainly being used to verify the results of analysis on the basis of local knowledge. More generally, the present review found that some Member States’ preliminary flood risk assessments would benefit from a clearer presentation of the methodologies applied to identify flood risk and assess its significance, for past and/or future floods. Nevertheless, in just under half the Member States, the criteria for identifying significant future floods have been updated based on current methodologies.

An assessment of the information reported on the impact of past floods was not included in the EU overview document (published in 2015) that discussed the Member States’ first ever preliminary flood risk assessments under the Floods Directive. This aspect was looked at this time around, also thanks to improved reporting requirements commonly agreed to with the Member States via the Common Implementation Strategy (a platform to cooperate for better implementation of the water acquis). The conclusion is that information on the impact of past floods is being collected, albeit variably. Some Member States do qualitative assessments, while others collect more detailed, quantitative, data. However, in 60% of river basins in the EU there are no data on the costs from flood damages. There is therefore room for improvement since collecting such data aids for example the calculation of costs and benefits and the prioritisation of measures. Considering other policy developments in the areas of disaster risk management and climate change, a more nuanced attention to disaster loss data is therefore strongly advisable.

For the vast majority of Member States there is some or strong evidence that the consequences of future flooding on human health, the environment, cultural heritage and economic activity are being considered. Attention to environment and cultural heritage appears to have risen compared to the past since the percentage of areas of potential significant flood risk where environment and cultural heritage were not found to be relevant dropped by around 10 percentage points.

Long term developments (socio-economic, infrastructure, land use) have been considered in most Member States but with varying degrees of rigour. There is also evidence that all Member States have considered climate change in their preliminary assessments; this is an improvement on the previous comparable review where the case was unclear for over a third of Member States.

2.Introduction and background

The Floods Directive 2007/60/EC (FD) came into force in 2007. It established a framework for flood risk management (FRM) and foresees 6-yearly cycles with the objective to reduce the risk of flood damage in the European Union (EU). The first cycle of implementing the FD covered the period 2010-15. The second cycle covers the period 2016-21. The latter is also the period of implementation of the first Flood Risk Management Plans (FRMPs), which were established by the Member States at the end of the first cycle. The first FRMPs 1 , but also the first cycle Preliminary Flood Risk Assessments (PFRAs), the Areas of Potential Significant Flood Risk (APSFRs) identified, and the Flood Hazard and Risk Maps (FHRMs) 2 were sequentially assessed by the European Commission (the Commission). During the second cycle, Member States are required to have reviewed and updated, by 22 December 2018, their first cycle Preliminary Flood Risk Assessments. This is the subject of the present document.

Member States report information to the Commission electronically using dedicated tools and databases developed under the European Environment Agency’s (EEA) Water Information System for Europe (WISE). The reporting guidance and relevant digital tools for reporting under the FD 3 have been updated for the second cycle and are available on the European Environment Information and Observation Network’s (EIONET) Common Data Repository (CDR) 4 .

The tables in this report have been generated from the data and information provided by the Member State. The electronic reporting format includes the requirement for the Member States to select from pre-defined options contained in lists (e.g. a list of criteria for identifying past floods with significant adverse impacts). The Member States selected the options that correspond to their respective situations when reporting to the Commission. In addition to the selection of options, the Member States also reported PFRA studies and internet links to further information and this information has also been evaluated as part of the assessment. This document reflects the situation as reported by the Member States to the Commission before the assessment and with reference to PFRAs prepared prior to the reporting. The situation in the Member States may have altered since then.

This document includes 26 of the 27 Member States 5 . The individual Member State assessment studies, published separately, provide the background to the present EU overview. 

Whereas a key role of the Commission is to check compliance with EU legislation, the Commission also seeks information to determine whether existing policies are adequate. The present EU overview and the individual Member States’ PFRA reviews conducted by the Commission, aim at strengthening flood risk management in the EU on the basis of good practice, as it emerges from the implementation of the FD by the Member States themselves. The Commission also collects information to create an EU-wide picture to inform the public on certain aspects of policy. The present document (and the individual reviews it is based on) therefore may take a broader perspective than the FD; the text of the latter being the only benchmark against which a Member States’ compliance to the FD should be judged.

3.Overview of timeliness and completeness of the information reported

Member States report information to the Commission electronically using dedicated tools and databases developed under the European Environment Agency’s (EEA) Water Information System for Europe (WISE). The information provision requirements included in the WISE/EIONET (European Environment Information and Observation Network) electronic reporting has been agreed with the Member States and is reflected in “Reporting Guidance” documents. The reporting guidance and relevant tools for reporting under the FD have been updated for the second cycle and are available on EIONET 6 . The information reported to WISE was the starting point for the assessment of Member States’ second cycle PFRAs. The majority of the statistics presented are based on information reported to WISE. Assuming that the Member States accurately transferred the information contained in their PFRAs to EIONET 7  and barring any undetected errors in the transfer of this information to WISE, arising from the use of interfacing electronic tools, these statistics should accurately reflect the content of the PFRAs.

3.1.Timeliness of the reported information

Table 1 shows the time periods over which the Member States reported information to the EIONET CDR. According to the timescales of the FD, information on the second PFRAs should have been reported by 22 March 2019. None of the Member States completed their reporting by March 2019 and only one third of Member States commenced reporting in March 2019. However, over half completed a substantial part of their reporting within a few months from March 2019. It should be noted however that due to an update in the folder structure of the reporting infrastructure and due to a later issue with the number of processes that the servers could handle, Member States were given until 30 August 2019 to complete their submissions. More generally, reasons for late reporting include one or more of the following: delayed preparation of PFRAs, data quality control issues or latent bugs in the reporting infrastructure, corrections and updates to previous submissions or provision of supplementary information.

As can been seen from Table 1, eight Member States 8 started uploading information in March 2019, but no Member States had completed their reporting by this date. Denmark and the Netherlands were the first Member States to complete reporting (in June 2019) followed by Slovakia in July 2019. By December 2020 all Member States, with the exception of Bulgaria, Cyprus, Greece, and Malta, had uploaded the bulk of the information required to allow the assessments to commence. It should be noted that some Member States updated the information reported after the assessments had commenced. By way of example only, Austria provided updated files in March 2021, and Latvia provided updated spatial data for APSFRs in January 2021. Greece, Cyprus and Malta reported by April 2021. At the time of writing, Bulgaria had not yet completed its reporting.

Table 1: Dates when Member States reports submitted into the EIONET CDR (for illustrative purposes only)

3.2.Completeness of the reported information

Table 2 shows the information reported by each Member State to the EIONET CDR. It should be noted that the FD does not require information about CAs or UoMs to be reported for every cycle if it has not changed. As a result, many Member States have not needed to report information on CAs or UoMs. This may also explain why some UoMs have not reported spatial data.

Twenty five of the 26 Member States that reported information presented evidence to show that a PFRA has been prepared for all UoMs. The exception is Spain where no PFRA was reported for the Balearic Islands (ES110) in time for the assessment 9 . The approach to preparing the PFRA varied between Member States. Those Member States with only one UoM 10 understandably prepared only one PFRA. Fourteen Member States 11 prepared one PFRA document that included all the UoMs in the Member States. Seven Member States 12 produced PFRAs for each UoM. In Finland, France and Portugal a nationally agreed template was used. In Italy, PFRAs were prepared for each River Basin District (RBD) covering several UoMs against a nationally agreed template.

The length and clarity of the PFRAs varied between the Member States. Some were clearly written and explained the methodology that has been used in way that would be easy for the general public to understand. Others were written in rather technical language that would be difficult for the layman to interpret, whilst others did not contain sufficient information to allow the adequacy of the methodologies employed to be determined. Some included hyperlinks to where more detailed methodological information could be found. Some included in-depth analysis of certain aspects of the PFRA, e.g. past floods 13 or climate change, but it was not always clear how this information had then been used in the assessment of flood risk. Most of the Member States published the PFRA online. One Member State had not made their PFRA available in this manner, whilst another had already consigned the documents to an archive server.

Table 2: Completeness of the information reported to EIONET Central Data Repository by Member States

Key:

Notes:

EE: No floods occurred in one UoM (EE3) therefore no APSFRs have been identified.

ES: No data was reported for UoM ES110 in time for the assessment.

HU: No change in spatial data for future floods since first cycle.

FI: No significant flood risk identified in two UoMs (FIVHA1 and FIWDA).

PL: No data reported for two UoMs (PL3000, PL4000). Incomplete data for PL8000 and PL9000.

4.UoMs and Competent Authorities

The FD provides that Member States may make changes to their administrative arrangements which include their Units of Management 15 (UoMs) and their relevant Competent Authorities (CAs). If such changes occur, Member States are required to notify the Commission within three months. None of the Member States have reported that they have made notable changes to administrative arrangements. The UoMs are shown in Figure 1.

Figure 1: Map of UoMs

Source: WISE, Eurostat (country borders)

5.Implementation of Article 4

5.1.Introduction to Article 4.2 of the Floods Directive

Article 4.2 of the FD requires Member States to undertake a preliminary flood risk assessment (PFRA). The Directive requires that PFRA be based on available or readily derivable information, such as records and studies on long term developments, in particular impacts of climate change on the occurrence of floods. The PFRA shall include at least the following:

·Maps of the river basin district at the appropriate scale including the borders of the river basins, sub-basins and, where existing, coastal areas, showing topography and land use (Article 4.2(a));

·A description of the floods which have occurred in the past and which had significant adverse impacts on human health, and for which the likelihood of similar future events is still relevant, including their flood extent and conveyance routes and an assessment of the adverse impacts they have entailed (Article 4.2(b));

·A description of the significant floods which have occurred in the past, where significant adverse consequences of similar future events might be envisaged (Article 4.2(c)); and

·Where the specific needs of the Member States require it, an assessment of the potential adverse consequences of future floods for human health, the environment, cultural heritage and economic activity, taking into account as far as possible issues such as the topography, the position of watercourses and their general hydrological and geomorphological characteristics, including floodplains as natural retention areas, the effectiveness of existing man-made flood defence infrastructures, the position of populated areas, areas of economic activity and long-term developments including the impact of climate change on the occurrence of floods (Article 4.2(d)).

Figure 2 shows the relationship between Articles 4.2 (b), (c) and (d), and Article 5 (the selection of APSFRs) and depicts the recommended steps in order to carry out a full Article 4 and Article 5 analysis.

Figure 2 (this and next pages): Flow charts showing the relationship between Article 4.2 (b), (c) and (d) (the PFRA) and the Article 5 (the selection of APSFRs)

5.2.Article 4.2(a) – Maps with topography and land use

As stated above, Article 4.2(a) states that the PFRA should include maps of the river basin district at the appropriate scale including the borders of the river basins, sub-basins and, where existing, coastal areas, showing topography and land use.

Figure 3 shows that all the Member States have included maps, or made them available through a map viewer.

Figure 4 shows the number of Member States that have presented strong evidence or some evidence 16 of the required features being included on the maps in their PFRAs and/or any interactive map viewers that had been made available to support the PFRA process. In some cases, the information required was shown on the map viewer, but not in the map published in the PFRA document, or vice versa, which accounts for the “some evidence” being noted. In the case of topography and land use “some evidence” has also been applied where only some elements of topography and land use have been included, or where the information has been included for some UoMs.

Table 3 shows which Member States have included which items in the PFRA. It can clearly be seen that whilst the borders of river basins are largely well represented in the maps, the same cannot be said for the borders of the sub-basins. Most Member States show some topographic and land use information, however, in several cases this could be improved (e.g. use a different scale) 17 . Twelve Member States provided links to specific flood related geoportals that allowed information directly related to the PFRA to be examined interactively 18 .

Figure 3: Number of Member States that have included maps in their PFRAs at an appropriate scale

Figure 4: Number of Member States that have included the required map features

Note: five Member States are landlocked (AT, CZ, HU, LU, SK) and therefore the display of coastal areas is not applicable.

Table 3: Information shown on each map by Member States

Key:

5.3.Article 4.2(b) – Assessment of past floods with significant impacts

Article 4.2(b) requires Member States to provide a description of past floods with known significant adverse impacts that may reoccur. Reporting requires a methodology for defining what constituted a ‘significant adverse impact’ at the time of flooding. To achieve this, Member States’ CAs need to collect information on the floods that occurred and the impacts that ensued.

5.3.1.General information on past floods

Figure 5 shows the total number of historic flood events 19 that were included by the Member States in the second cycle reporting exercise, whilst Figure 6 shows the past floods that were reported as having occurred during the period 2012 – 2018 (i.e. during the second cycle); in total around 2 700 flood events. Three Member States 20 did not report any historic flood information to the EIONET CDR. However, in their PFRA reports:

·Lithuania provided information on 17 significant flood events that occurred in the period 2011-2017;

·Malta provided information on eight pluvial foods that had occurred during the second cycle.

·Slovenia provided information on a total of 360 flood events of which 145 occurred after 2010.

Figure 5: Total number of historic flood events (predating 2012 included) as reported to the EIONET CDR in the second cycle by Member States 21

Figure 6: Number of reported flood events that occurred in the period 2012 – 2018 as reported to the EIONET CDR in the second cycle by Member States 22

Croatia reported the largest number of events (699, slightly over a quarter of the total number of events reported by Member States between 2012 and 2018).

Figure 7 summarises the time periods of the floods reported in the second cycle (floods that occurred during the second cycle, 2012-18, but also before). This shows that half of the floods reported in the second cycle relate to time periods before 2012. The oldest flood event reported (by Poland) in the second cycle was from 1829. Slightly over one third of the flood events reported relate to the period 2000-2010 23 .

Figure 7: Time periods of flooding as reported to the EIONET CDR in the second cycle

5.3.2.Sources, mechanisms and characteristics of past floods

Figure 8 shows the sources of flooding for the flood events reported that occurred in the period 2012-2018. This shows that for most Member States fluvial flooding remains the most significant source, although for Sweden, pluvial flooding and seawater flooding are reported as the only sources of floods in this period. Czechia, Denmark, Estonia, Greece, Finland, Hungary and Latvia have not reported pluvial flooding as a source of past floods in the period 2012-2018. In addition to Sweden, floods from seawater have been reported by 13 other Member States 24 . Floods from groundwater have been reported by four Member States 25 , whilst floods from artificial manmade infrastructure have been reported by six Member States 26 .

Figure 8: Sources of flooding for the flood events that occurred in the period 2012-2018 as reported to the EIONET CDR in the second cycle by Member States

Notes:

-More than one source of flooding may have been identified for a flood event.

-Bulgaria did not report in time, Lithuania, Malta and Slovenia did not report flood event data.

Figure 9 shows the sources, mechanisms and characteristics 27 of flood events that took place in the period 2012-2018. This shows that at an EU level fluvial flooding is the most significant source (99% of flood events 28 ), followed by flooding from seawater (slightly under six for every ten floods) and pluvial flooding (45%). Nearly all floods were generated from natural exceedances (97%) with blockages, defence failures and defence exceedance being other significant causes. The most common characteristics were flash flooding (slightly over two thirds), medium onset flooding (one third), other rapid onset flooding (slightly under three out of every 10) and slow onset flooding (slightly over a quarter). It should be noted that although the source of flooding was reported as “no data” for only 2% of events the mechanism of flooding was reported as no data for slightly over a third of flood events, and the characteristics of flooding were reported as no data for slightly over a quarter of flood events.

Figure 9: Source, mechanisms and characteristics of flood events occurring in the period 2012-2018 as reported to the EIONET CDR in the second cycle

Note: More than one source, mechanism and characteristic may have been identified for a flood event

5.3.3.Extent and frequency or recurrence of past floods

Some Member States included detailed information in their PFRAs on how they have collected information on past floods. For example, Belgium, the Republic of Ireland and Portugal have developed standardised templates, on-line data collection tools and have drawn on wider information sources such as newspapers 29 . Some Member States have cast the net more widely in respect of the organisations from whom information is collected, for example, Poland supplemented information on the floods that took place before 2012 (so during the first cycle) with new data obtained from various sources including a survey of municipalities, regional authorities, fire brigades and other stakeholders.

The amount and quality of quantitative information reported on the duration, extent and frequency of past floods varies widely between Member States. All Member States that reported past floods reported the date of the flooding. Figure 10 shows the quantitative data reported for date, location, extent, duration, and frequency/recurrence at the event level by Member States.

Figure 10: Quantitative data reported to the EIONET CDR in the second cycle on the extent, duration and frequency/recurrence of flood events

Note: Bulgaria did not report, Lithuania, Malta and Slovenia did not report flood event data.

Table 4 shows the number of UoMs in each Member States for which data on extent, duration and frequency/recurrence was reported.

Table 4: Member States which reported quantitative data on the extent, duration and frequency or recurrence of flood events (for all historic events reported)

Notes:

AT: Data reported for AT1000 only. No floods occurred in AT2000 or AT5000.

CZ: Data reported for CZ5000 only. No floods occurred in CZ1000 or CZ6000.

DE: Floods reported for 7 UoMs. No floods occurred in DE7000, DE9500, or DE9610.

EE: Data reported for EE1 and EE2. No floods occurred in EE3.

ES: Data reported for 21 UoMs. No floods occurred in ES014, ES150, or ES160.

FI: Data reported for FIVHA3, FIVHA4, FIVHA5. No floods occurred in FIVHA1, FIVHA2, FIVHA6, FIVHA7, or FIWDA.

FR: Data reported for 13 UoMs. No floods occurred in FRB2.

IT: Floods reported for 40 UoMs. No floods occurred in ITI022, ITI029, ITN004, ITR061, ITR151, ITR152, or ITR154.

LU: Floods reported in LU RB_000 only. No floods occurred in LU RB_001.

PL: Floods reported in 6 UoMs. No floods occurred in PL3000, PL4000 or PL8000.

Key:

All Member States that reported past floods reported the date of flooding for all events, and all but two (Spain and Poland) reported the location for all events. Only Romania reported data on the extent, duration and frequency/recurrence for all flood events 30 . Germany and Latvia reported all this information for some events, but not for all. Only six Member States 31 reported some information on the extent of flooding. As stated above, Romania reported the extent of flooding for all events, whilst Poland reported this information for 96% of events and Latvia for 86% of events. Quantitative information on the extent of flooding (either area inundated, or river length flooded) was not reported in nearly eight out of every ten UoMs. Information on the duration of flooding was reported by most Member States with ten Member States 32 reporting duration for all flood events. Of those Member States that reported flood events, three Member States 33 did not report information on the duration of flooding for any events. Nine Member States 34 reported information on the frequency or recurrence of all flood events, but of those Member States that reported flood events four Member States 35 did not report any information. Also the findings of this paragraph point towards increasing the effort of recording information around flood events in order to prepare better responses in the future.

In the first cycle, four Member States 36  applied Article 13.1(b) across all their UoMs and were not required to report information on historic flood events. Other Member States reported flood events with data on type and consequences. At the time the first cycle EU overview document 37 was written, a total of 18 153 historic flood events were reported: 15 660 with data, 2 493 with no data. However, the assessment did not make a distinction between data on the extent, duration and frequency of flooding and data on the impacts of flooding. 

The amount of quantitative information reported on the extent, duration and frequency of past events has improved in the second cycle, but there is scope for further improvements in the third cycle of reporting.

5.3.4.Quantitative data on impacts of past floods

In the 2020 national reporting of risk assessments 38 under the Union Civil Protection Mechanism (UCPM) 39 , floods were the most commonly identified natural disaster of concern, and floods risk management is therefore an important component of overall disaster risk management. Considering the effects of climate change, it is expected this will continue being the case.

Being in possession of robust disaster loss data improves modelling of disaster risk, the calculation of cost and benefit ratios to ensure effective and transparent investment decisions (including the prioritisation of measures) and helps the public understand the importance of the investments. Also, the assessment of the overall economic damage from disasters underpins the understanding of the macroeconomic impacts of disasters for the purpose of managing public finances, monetary stability and the resilience of financial systems 40 .

To appreciate the amount of future potential losses, PESETA IV 41 projected that with 3°C global warming, river flood damage in the EU and UK in 2100 would be six times larger than current losses, reaching €48 billion/year. Without mitigation and adaptation measures, annual damage from coastal flooding in the EU and UK could increase sharply from 1.4 €billion nowadays to almost 240 €billion by 2100.

The amended UCPM 42  therefore foresees (Article 6) that “…MS shall:… (f) in line with international commitments, improve the collection of disaster loss data at national or the appropriate sub-national level to ensure evidence-based scenario building…”

Further, the EU is party to the UN Sendai Framework for Disaster Risk Reduction, which requires the evaluation of disaster-related losses and economic impacts, and sets a target to reduce such losses by 2030 43 . At EU policy level, climate-related disaster loss data is needed for several policy areas. For example, to improve the economic foundation of adaptation policy and disaster management planning, to increase transparency about climate risks, to inform the European Semester, or to tailor Common Agricultural Policy support for loss recovery and prevention 44 .

Thus, there is a need to improve the gathering and access to disaster loss data. As a response to this need, the Risk Data Hub 45 was developed by the Commission and hosted in the Disaster Risk Management Knowledge Centre. The Risk Data Hub is a geoportal offering EU wide harmonized multi-hazard risk and loss data. It is a central repository for recording, sharing and monitoring curated disaster damage and loss data obtained from various open source databases. The new EU Strategy on Adaptation to Climate Change 46  promotes and supports the use of its Risk Data Hub to harmonise the recording and collection of comprehensive and granular climate-related risk and losses data. It also encourages national level public private partnerships to collect and share such data.

The FD introduced in 2007 the requirement for Member States, on the basis of available or readily derivable information, to describe past floods and assess their adverse impacts, and to make an assessment of the potential adverse consequences of future floods. The data on past floods collected via the PFRA process could make a useful contribution to closing loss data gaps. Consequently, Member States were asked, through the updated reporting infrastructure 47 , to provide more detailed information, where available, on the costs of damage resulting from historic flood events as:

·The total cost of damage in €;

·The total cost of damage as a proportion of GDP; or

·A qualitative assessment as a damage class:

­I - Insignificant

­L - Low

­M - Medium

­H - High

­VH - Very high (VH)

­NA – Not Applicable; or

­U – Unknown.

Figure 11 shows the types of impact information that was provided at flood event level by Member States. Of those Member States who reported event data, nine 48 did not report a qualitative or quantitative assessment of the level of impact of flood events, although the four broad types of impact as required by the FD (economy, health, cultural heritage, environment) were identified.

Figure 11 Data reported on the impact of flood events at event level as reported to the EIONET CDR in the second cycle by Member States for all historic floods reported 49

Notes:

-Bulgaria did not report, Lithuania, Malta and Slovenia did not report flood event data.

-Although it appears Hungary provided quantitative data on the impact of all flood events reported, it should be noted that Hungary only reported one event in the second cycle.

Six other Member States 50 reported quantitative data totalling €3 778 527 772 for historic floods (over the years 2012-18), with Italy doing so for 85% of a total of 310 flood events. The average level of damage per event for these six Member States is shown in Figure 12.

Figure 12: Average damage per flood event for events where damage was reported as reported to the EIONET CDR in the second cycle

Note: Greece also provided data on the damage by event, but the data extracted from the CDR appeared to be erroneous.

Eleven Member States 51 provided a qualitative assessment of the impact of flood events, with Cyprus and Luxembourg doing so for all flood events, and Austria doing so for 98% of events.

The data reported on the impact of past floods in the first cycle is described in the section on duration, extent and frequency of flooding above. The amount of quantitative information reported on the impact of past events has improved in the second cycle, but there is scope for further improvements in the third cycle of reporting.

In addition to the data reported to the EIONET CDR on extent, duration, recurrence and impact of flooding, there is often more detail presented in the PFRA reports or in other documents. For example, Austria provided a file containing detailed information on each of the 45 significant floods that have occurred since 2011. The information collected includes the date and location of each flood, the duration, frequency, origin, cause and mechanisms of each flood and the damage caused in terms of the area, inhabitants affected and total damage costs. Information is also included on the cost of damage prevention in the future as a total cost per event ranging from €20k-€5 million (total cost €5.9 million), and the costs incurred to repair damage per event by one of the two Federal Agencies responsible, the Federal Hydraulic Engineering Administration 52  (cost ranged from €2 - €5.5 million and totalled in excess of €18.25 million). The spreadsheet allows for collection of data from the other Federal Agency with responsibility, the Torrent and Avalanche Control 53 , but no data has currently been included.

Hungary, who reported only one significant flood in the second cycle PFRA, provided a detailed textual description in the PFRA document 54 , in addition to providing qualitative and quantitative information (economic damage) in the report to the EIONET CDR 55 .

5.3.5.Methodologies and criteria used for the assessment of the significance of the impacts of past floods

Article 4.2(b) of the FD requires Member States to identify the adverse impacts on human health, the environment, cultural heritage and economic activity of past floods for which the likelihood of similar future events is still relevant, including their flood extent, and conveyance routes and an assessment of the adverse impacts they have entailed.

Figure 13 shows the criteria used in UoMs to assess the impact of past floods, and Table 5 shows the criteria used in the UoMs in each Member State. The most used criteria 56 are:

·Return period (nearly two thirds of UoMs)

·Flooded area (slightly over six for every 10 UoMs);

·Residents affected (exactly six out of every 10 UoMs);

·Commercial area affected (57% of UoMs); and

·Buildings affected (54% of UoMs).

A number of UoMs (28%) used other criteria. Examples of these include the number of fatalities caused by flooding, trams being blocked for several days and ecological damage due to pollution of a protected area (Austria), the number of “memorable” events and the number of fatalities (France), the declaration of emergency by the state and/or fatalities due to floods (Greece) and the number of evacuations (the Netherlands). Expert judgement was used by 17% of UoMs. Examples of how this was applied include:

·The use of local knowledge and field expertise to identify the floods with the most significant impacts (Belgium/Wallonia).

·The historical storm surges are assessed and described with five categories: Water Level, Meteorology, Flood, Damage and Human Consequences. For each category, the parameters Data Availability, Data Quality and Phenomenon were assessed on a scale from 0 to 3. Under these, the categories Injuries and People represent the consequences of the flood, where the parameter Phenomenon describes the extent. If the phenomenon of an incident is 2 or greater, the flood is defined as having extensive damage (Denmark).

·Assessment of impacts conducted by gathering all of the available data and reviewing it by experts on case to case basis (Lithuania).

·Definition of areas where floods with significant adverse impacts may reoccur based on size and location of the sub-catchment, predominant land use of the sub-catchment, runoff volume generated at the basin outlet of the sub-catchment, documented flood events in the past within the sub-catchment (Malta).

·Data on historical floods obtained and updated mainly by competent units as a result of a survey of municipalities, communes, provinces, crisis management centers, provincial fire brigade units, irrigation and water authorities (Poland).

Figure 13: Criteria used by UoMs for the assessment of the impact of past floods as reported to the EIONET CDR in the second cycle

Note: More than one criteria can be used to assess the impact of past floods.

Table 5: Criteria used in each Member States to assess the impact of past floods according to Article 4.2(b)

Key:

Figure 14 shows the number of Member States where the results of the individual Member States assessments show that there is strong, some or no evidence of a clear methodology being in place for the assessment of past floods. This shows that almost two thirds of the Member States have presented strong evidence to show that a clear methodology is in place for the assessment of past floods 57 . Two Member States (Malta and Slovakia) presented no evidence in the reported information.

Figure 14: The number of Member States where there is strong, some, or no evidence of a clear methodology being in place for the defining past floods with significant adverse impacts as required by Article 4.2(b)

Figure 15 shows the percentage of UoMs (as a proportion of the total of 206 UoMs in the EU) that indicated impacts of past flood events in the second cycle PFRAs. The greatest number of UoMs (nearly four fifhts) reported economic impacts on infrastructure as a result of flooding and slightly over two thirds reported economic impacts on property. Impacts on human health as a result of flood events were reported by 56% of UoMs, and impact on the health and social well-being of communities reported by slightly under half of UoMs. The impact of flooding on cultural heritage has been accounted for with 46% of the UoMs reporting impacts of flood events on cultural assets but the impact of flooding on the environment appears to be less prevalent with a bit over a quarter of UoMs reporting impacts to protected areas, just under a quarter reporting impacts to water bodies and a bit less than a quarter reporting impacts on pollution sources.

Figure 15: Percentage of UoMs that have reported impacts of past floods to the EIONET CDR in the second cycle

Figure 16 presents the impact reported for flood locations for events that occurred only in the period 2012-2018. “Not applicable” was reported as the most significant impact for impacts on human health (84%), the environment and cultural heritage (each 82%). However, impacts on infrastructure (assets such as utilities, power generation, transport, storage and communication) were reported in 43% of locations, on property (such as homes and businesses) in 29% of locations and on economic activity (such as manufacturing, construction, retail, services and other sources of employment) in 12% of locations. Other economic impacts were reported in 28% of locations, these included:

·Economic activity which is significant in terms of ensuring the functions vital to society (Finland);

·Hydraulic works-longitudinal defence works such as embankments, bank walls, groynes (Italy).

Figure 16: Number of flood locations 58 in the period 2012 – 2018 that have been identified as having been impacted as reported to the EIONET CDR in the second cycle

Many of the UoMs use quantitative data to evaluate the significance of the impact of past floods. For example, in the Netherlands 59 , the impact on populations and the cost of the damage incurred is assessed, whilst in Latvia 60 the focus is on the expenditure that has been required to make good the damage. In Belgium (Flanders), the assessment of past floods is based on data from the disaster fund and the insurance sector.

The complexity of the methodology used to identify significant past floods varies widely between Member States. Some use a simple methodology, for example Slovenia 61 identified past floods as significant if: (1) there were fatalities; (2) there was damage to people’s property; (3) there was damage to infrastructure including cultural heritage. Others, such as Portugal, applied a classification scheme based on the damage to a number of receptors which were combined to give an overall classification 62 .

The two attributes that are singled out for consideration in the assessment of significance of impact according to Article 4.2 (b) are:

·The extent of past flooding; and

·Conveyance routes.

Figure 17 presents the number of Member States where the results of the Commission’s individual Member State assessments show that there is strong, some or no evidence of either the extent of past flooding and conveyance routes being taken into consideration in the assessment of impact of past floods.

For 11 Member States 63 there is strong evidence that the extent of flooding has been considered, with some evidence presented for a further eight Member States 64 . For example, in Denmark, extent of flooding is classified on a scale of 0, 1, 2 or 3 depending on data availability and quality 65 . Italy has developed a database (FloodCat) and a data entry platform (Heroic) which enables information on flood events, including spatial data on the extent of flooding, to be captured and used for the assessment of past floods as required by both Article 4.2(b) and 4.2(c). 66  

However, according to the reported information, only six Member States 67 presented strong evidence of having considered conveyance routes in the assessment of past floods. Luxembourg provided a detailed assessment of a flood that occurred in January 2011 which included the use of satellite imagery to map the conveyance route of flood which will contribute to the refinement of models in the future 68 . A further nine 69  Member States presented some evidence of having done so. Some Member States noted that consideration of conveyance routes is an implicit part of the PFRA (e.g. Austria) or is part of flood hazard modelling and mapping (e.g. Sweden).

Figure 17: Number of Member States that have provided evidence that demonstrates whether the extent of past flooding and conveyance routes have been considered in the assessment of past floods according to Article 4.2(b) 70

5.4.Article 4.2(c) – Assessment of past floods without known significant impacts that may have significant impacts if repeated in the future

5.4.1.Methodologies and criteria used for the assessment of significant past floods without known impacts

Article 4.2(c) requires Member States to include a description of the significant floods (in terms of extent) which have occurred in the past (without significant impacts however), where significant adverse consequences of similar future events might be envisaged due to climate and/or socio-economic change.

Figure 18 shows the number of Member States where strong evidence, some evidence or no evidence was found of a clear and distinct methodology being in place for the implementation of Article 4.2(c). Only seven Member States 71  were found to have provided strong evidence of a clear and distinct methodology being in place for the assessment of such floods whilst 12 Member States 72 presented some evidence. In most of these cases there was evidence that this had been considered, but the methodology was not presented in a clear way, or there was no distinction between the methodologies applied in relation to Article 4.2(b) or Article 4.2(d). The remaining Member States presented no evidence that Article 4.2(c) had been addressed. In most cases no reasons were provided for this, but Poland did state that Article 4.2(c) had not been applied due to a lack of data.

Figure 18: Number of Member States where there is strong, some or no evidence of a clear methodology being in place for the assessment under Article 4.2(c)

Figure 19 shows the criteria used by UoMs to define significant past floods (without impacts at the time) which may have an impact in the future as reported by Member States to the EIONET CDR.

Figure 19: Criteria for defining the significance of past floods without known significant adverse impacts under Article 4.2(c) as reported to the EIONET CDR in the second cycle.

Note: More than one criteria can be applied for the assessment of past floods.

Table 6 shows which of these criteria were selected by which Member States. The most frequently used criteria were flooded area and return period (slightly under two thirds of UoMs each), buildings affected (also a bit under two thirds of UoMs), residents affected, and commercial area affected (61% of UoMs each). Weighting systems were used by only 14% of UoMs. Slightly less than a fifth of UoMs reported that other criteria were used, these included (similar to Article 4.2(b)):

·Number of fatalities caused by flooding, trams being blocked for several days, ecological damage due to pollution of a protected area (Austria);

·Harmful consequence for the environment and cultural heritage (Finland);

·Number of deaths and “memorable” events (France);

·Indication of frequency, number of evacuations and date of occurrence, after 1900 (the Netherlands).

Expert judgement was used in 28% of UoMs, mostly in conjunction with other criteria. Examples of how this was applied include:

·Definition of areas where significant past floods without known significant adverse impacts but where significant adverse consequences might be envisaged based on size and location of the sub-catchment, predominant land use of the sub-catchment, runoff volume generated at the basin outlet of the sub-catchment, documented flood events in the past within the sub-catchment (Malta).

·The likely impact that would occur should urban development occur in areas that have flooded in the past with no impact (Sweden).

Only Croatia used expert judgement alone. Four floods were identified for further assessment but the exact methodology used is not clear.

In the first cycle’s EU overview document no distinction was drawn between the methodology for Articles 4.2(b) and 4.2(c). Although in the second cycle this has become clearer, it is not possible to compare and determine whether the situation with respect to discerning between Articles 4.2(b) and 4.2(c) has improved in the second cycle. Member States should consider providing clearer information on how Article 4.2(c) has been applied. A comparison will be possible in the third cycle.

Table 6: Criteria used by each Member States for defining past floods without known significant adverse impacts under Article 4.2(c)