The flood risk estimates how often a river overflowing brings at least 0.5 m of water to the site, taking local flood defences into account. It uses the WRI Aqueduct river flood maps (cells of about 1 km, return periods from 2 to 1,000 years): the more often the water arrives, the higher the risk. The maps model rivers with large catchments: small streams and channels and flooding from rain water are not included, so class A means no flood of a large river in the maps, not the absence of any water risk. Heavy rain has its own risk. In Switzerland, in the 15 cantons that publish their hazard maps with a licence for commercial use, the class also takes them into account, and they include small streams and channels: the higher of Aqueduct and the cantonal map counts. In Italy the class also takes into account the ISPRA national mosaic of the flood hazard areas of the flood risk management plans: the higher of Aqueduct and the mosaic counts. In Austria the class also takes into account the official flood areas of the Floods Directive (30, 100 and 300 years): the higher of Aqueduct and the map counts. In England the class also takes into account the Environment Agency's Risk of Flooding from Rivers and Sea map, with the likelihood of water reaching at least 0.3 m: the higher of Aqueduct and the map counts.
Indicator. Return period (years) of the first flood that brings 0.5 m of water to the site, taking the local flood defences into account. It is the same basis as the euro damage.
Data. WRI Aqueduct Floods v2 river flood maps (cells of about 1 km, 9 return periods from 2 to 1,000 years, present climate, simulated without defences) and the FLOPROS protection standards (Scussolini et al. 2016). The extreme rainfall indices Rx1day and Rx5day from CMIP6 (Copernicus CDS) remain as indicators: return period of the models' value against 1850–1970.
Limit: Aqueduct, like the JRC global maps, models only rivers with large catchments (hundreds of km²); small streams and channels and flooding from rain water are not included, so class A does not rule out every water risk at the site.
In Switzerland the cantonal hazard maps are also used (federal model MGDM 166.1 v1.3, downloaded from geodienste.ch), from the 15 cantons that publish them with a licence for commercial use and with intensities per scenario: AG, AR, BE, BL, GL, JU, LU, SH, SO, SZ, TG, TI, VS, ZG, ZH. AI, FR and GE publish only hazard degrees, which do not say how often the water arrives, and are not used. The cantonal maps include small streams and channels, but only in the areas the canton assessed (usually built-up areas).
In Italy the ISPRA 2020 national mosaic of the flood hazard areas is also used (IdroGEO platform, [CC BY 4.0](CC BY 4.0) licence), which joins the maps of the districts' flood risk management plans: high (P3), medium (P2) and low (P1) hazard scenarios. It gives the extent of the areas only, without depth: any flooded area counts as damaging. Methods and the way defences are considered change from district to district, as ISPRA states. In 2020 the P3 scenario was missing for the territory of the former regional basin authority of the Marche, as ISPRA states: there frequently flooded areas appear only from P2 (100 years).
In Austria the national flood areas reported under the Floods Directive in 2020 are also used (ministry BMLUK and Umweltbundesamt, INSPIRE service, [CC BY 4.0](CC BY 4.0) licence): scenarios of 30 years (HQ30), 100 (HQ100) and 300 (HQ300). They join discharge studies, hazard zone plans and the HORA zoning; the ministry's guideline treats HORA as less precise data, and where a study is missing some areas are derived in a simplified way (for example the 300-year area by widening the 100-year one by 25 m). The 300-year scenario also includes defences being overtopped or failing. They give the extent of the areas only, without depth: any flooded area counts as damaging.
In England the Environment Agency's Risk of Flooding from Rivers and Sea map is also used (2 m grid, [Open Government Licence v3.0](Open Government Licence v3.0)): for each cell the yearly likelihood of river and sea water reaching at least 0.2, 0.3, 0.6, 0.9 and 1.2 m, in four categories: high (1 in 30 or more), medium (1 in 30 to 1 in 100), low (1 in 100 to 1 in 1,000) and very low (less than 1 in 1,000). It takes into account the defences and the chance of them being overtopped or failing. Where the Environment Agency kept the previous version (50 m grid) there are no depths: there the class reads the category of any flooding (the primary map) and no 0.6 m value is shown. Wales, Scotland and Northern Ireland are not included. Contains public sector information licensed under the Open Government Licence v3.0. © Environment Agency copyright and/or database right 2025. All rights reserved.
Calculation.
(1) The water depth at the site is read in the 9 maps.
(2) The return period at which the water reaches 0.5 m is found, interpolating between two maps on the logarithm of the return period (an xNatura choice: ThinkHazard! reads only the 10, 50 and 1,000-year maps).
(3) A flood smaller than the defences does not arrive: the return period used is the larger of this one and the protection standard of the region.
(4) The class is the same in every scenario and year, because the maps are of the present climate.
(5) Where there is an official map, the first return period at which the map brings damaging water to the site's point is read. The class is the higher of the Aqueduct one and the map one: the map can only raise it. The official maps already take the existing defences into account, each in its own way (see Data), so FLOPROS is not applied.
In Switzerland a flood of medium or strong intensity counts: at least 0.5 m of water, or velocity times depth of at least 0.5 m²/s (Swiss intensity classes, Loat and Petrascheck 1997), the same 0.5 m threshold as ThinkHazard!. Map scenarios: from 5 to 300 years; the extreme scenario counts as 1,000 years. Valais publishes all water processes together (debris floods and bank erosion included) and they are used as published. Almost every canton publishes only the 30, 100 and 300-year scenarios (Ticino also 5, 10, 20, 50 and 500): there 30 means “30 years or more often”, and the map alone reaches class C at most. The map is read on a 5 m grid where a flood area counts in every cell it touches, because torrent strips are often narrower than 5 m (an xNatura choice).
In Italy P3 counts as 20 years and P2 as 100, the lower bounds of Legislative Decree 49/2010 (20 to 50 and 100 to 200 years). For P1 the decree gives no return period (“low probability of floods or extreme event scenarios”) and the districts use from just over 200 to 500 years (over 200 for the Arno and the Tuscan basins, 300 and 500 in the Po district, 300 in the Eastern Alps); 300 years are used (an xNatura choice): any value between 50 and 1,000 years gives the same class B. For P3 and P2 too some districts use return periods other than the decree's (P3 from 2 to 10 years in some units of the Po district: Po, Reno, Romagna basins; P2 from 30 years in Tuscany): there the Italian map can give a lower class than the plan. On its own the Italian map reaches class D at most.
In Austria each scenario counts with its own return period: HQ30 30 years, HQ100 100, HQ300 300; on its own the Austrian map reaches class C at most.
In England the class reads the category of water reaching at least 0.3 m, the published depth just below the 0.5 m of Aqueduct and ThinkHazard!: so the map never says a site is safer than the 0.5 m threshold would (an xNatura choice). Each category counts with its longest return period: high 30 years, medium 100, low 1,000; so 30 means “30 years or more often” and the map alone reaches class C at most. The very low category does not change the class. The same reading for 0.6 m is shown next to it.
The maps of Italy, Austria and England are read on a 10 m grid where every cell touched by an area counts.
Thresholds → level A–F.
ThinkHazard! (GFDRR) river flood hazard levels: high if 0.5 m of water arrives within 10 years, medium within 50, low within 1,000, very low beyond; translated into the six classes as for gusts and snow:
• F: at most every 3.2 years
• E: 3.2–10 years
• D: 10–22.4
• C: 22.4–50
• B: 50–1,000
• A: beyond 1,000 years, or defences of at least 1,000 years.
If the maps cannot be read, the class falls back to the rainfall indices (previous version).
From physical to impact. The physical level (A–F) is combined with the production process's sensitivity through the risk matrix (6 physical levels by 5 sensitivity bands) → impact level. The euro amount of this risk is the one described in the "Economic damage" paragraph (Version 2026): the damage factor per class of Version 2025 is no longer used for this risk.
Economic damage. Besides the class, the loss in euro is estimated as a share of the asset value (book value), with the WRI Aqueduct Floods method (Ward et al. 2020).
• Water depth: WRI Aqueduct Floods v2 river flood maps, cells of about 1 km, for 9 return periods (2, 5, 10, 25, 50, 100, 250, 500 and 1,000 years), simulated without defences. They are maps of the present climate, used for every scenario and year.
• Damage: JRC global depth-damage curves (Huizinga et al. 2017) for industrial and commercial buildings, those of the site's continent (the global curve where the continent has none), multiplied by 0.6: the JRC treats 40% of a building as not damageable by water, and here the asset value is the book value. No water, no damage. Industrial or commercial use from the ISIC sections of the site's production processes (sections A to F, that is agriculture, mining, manufacturing, energy, water and waste, construction: industrial; G to U, services: commercial; a mixed site in proportion to its processes); without processes, half and half.
• Protection: floods smaller than the protection standard of the region do no damage. FLOPROS standard (Scussolini et al. 2016): design level, else policy level, else modelled level; the Netherlands 1,000 years; where there is no value, 2 years, that is no defences.
• Expected annual loss: area under the curve of the loss against the exceedance probability, only for the floods beyond the protection level and up to the 1,000-year flood. Loss at the 100-year flood: 0 if the protection is at least 100 years. Basis: published source.
The class and the euro come from the same maps and the same defences, except where an official map raises the class: the euro stays the Aqueduct one (the depths of the official maps are not used for the euro yet), so in that case the class can be high with zero loss. The variable “Source of the flood class” says which map set the class.
Sources.
• [Copernicus Climate Data Store (CDS)](Copernicus Climate Data Store (CDS))
• [Indices for monitoring changes in extremes based on daily temperature and precipitation data](Indices for monitoring changes in extremes based on daily temperature and precipitation data)
• [ETCCDI Climate Change Indices](ETCCDI Climate Change Indices)
• [Handbook of Drought and Flood Indices](Handbook of Drought and Flood Indices)
• [Guidelines on Analysis of Extremes in a Changing Climate in Support of Adaptation to Climate Change](Guidelines on Analysis of Extremes in a Changing Climate in Support of Adaptation to Climate Change)
• [ThinkHazard! methodology report (GFDRR)](ThinkHazard! methodology report (GFDRR))
• [Ward et al. 2020, Aqueduct Floods methodology](Ward et al. 2020, Aqueduct Floods methodology)
• [Huizinga et al. 2017, Global flood depth-damage functions](Huizinga et al. 2017, Global flood depth-damage functions)
• [Scussolini et al. 2016, FLOPROS: an evolving global database of flood protection standards](Scussolini et al. 2016, FLOPROS: an evolving global database of flood protection standards)
• [Cantonal hazard maps, geodienste.ch (MGDM 166.1)](Cantonal hazard maps, geodienste.ch (MGDM 166.1))
• [Loat and Petrascheck 1997, Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten](Loat and Petrascheck 1997, Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten)
• [ISPRA (2020) Pericolosità e indicatori di rischio alluvioni, piattaforma IdroGEO](ISPRA (2020) Pericolosità e indicatori di rischio alluvioni, piattaforma IdroGEO)
• [Autorità di bacino distrettuale del fiume Po (2020), Relazione metodologica delle mappe di pericolosità e rischio di alluvione](Autorità di bacino distrettuale del fiume Po (2020), Relazione metodologica delle mappe di pericolosità e rischio di alluvione)
• [Autorità di bacino distrettuale delle Alpi Orientali, Piano di gestione del rischio di alluvioni 2021-2027](Autorità di bacino distrettuale delle Alpi Orientali, Piano di gestione del rischio di alluvioni 2021-2027)
• [Autorità di bacino distrettuale dell'Appennino Settentrionale, Relazione metodologica delle mappe di pericolosità e rischio di alluvione (II ciclo)](Autorità di bacino distrettuale dell'Appennino Settentrionale, Relazione metodologica delle mappe di pericolosità e rischio di alluvione (II ciclo))
• [BMLUK, Hochwassergefahren- und Hochwasserrisikokarten (Hochwasserrichtlinie)](BMLUK, Hochwassergefahren- und Hochwasserrisikokarten (Hochwasserrichtlinie))
• [BMLUK (2014) Hochwasser-Gefahrenkarten, fachlicher Leitfaden](BMLUK (2014) Hochwasser-Gefahrenkarten, fachlicher Leitfaden)
• [Environment Agency (2025), Risk of Flooding from Rivers and Sea](Environment Agency (2025), Risk of Flooding from Rivers and Sea)
• [Open Government Licence v3.0](Open Government Licence v3.0).
Indicator. Return period (years) of the first flood that brings 0.5 m of water to the site, taking the local flood defences into account. It is the same basis as the euro damage.
Data. WRI Aqueduct Floods v2 river flood maps (cells of about 1 km, 9 return periods from 2 to 1,000 years, present climate, simulated without defences) and the FLOPROS protection standards (Scussolini et al. 2016). The extreme rainfall indices Rx1day and Rx5day from CMIP6 (Copernicus CDS) remain as indicators: return period of the models' value against 1850–1970.
Limit: Aqueduct, like the JRC global maps, models only rivers with large catchments (hundreds of km²); small streams and channels and flooding from rain water are not included, so class A does not rule out every water risk at the site.
In Switzerland the cantonal hazard maps are also used (federal model MGDM 166.1 v1.3, downloaded from geodienste.ch), from the 15 cantons that publish them with a licence for commercial use and with intensities per scenario: AG, AR, BE, BL, GL, JU, LU, SH, SO, SZ, TG, TI, VS, ZG, ZH. AI, FR and GE publish only hazard degrees, which do not say how often the water arrives, and are not used. The cantonal maps include small streams and channels, but only in the areas the canton assessed (usually built-up areas).
In Italy the ISPRA 2020 national mosaic of the flood hazard areas is also used (IdroGEO platform, [CC BY 4.0](CC BY 4.0) licence), which joins the maps of the districts' flood risk management plans: high (P3), medium (P2) and low (P1) hazard scenarios. It gives the extent of the areas only, without depth: any flooded area counts as damaging. Methods and the way defences are considered change from district to district, as ISPRA states. In 2020 the P3 scenario was missing for the territory of the former regional basin authority of the Marche, as ISPRA states: there frequently flooded areas appear only from P2 (100 years).
In Austria the national flood areas reported under the Floods Directive in 2020 are also used (ministry BMLUK and Umweltbundesamt, INSPIRE service, [CC BY 4.0](CC BY 4.0) licence): scenarios of 30 years (HQ30), 100 (HQ100) and 300 (HQ300). They join discharge studies, hazard zone plans and the HORA zoning; the ministry's guideline treats HORA as less precise data, and where a study is missing some areas are derived in a simplified way (for example the 300-year area by widening the 100-year one by 25 m). The 300-year scenario also includes defences being overtopped or failing. They give the extent of the areas only, without depth: any flooded area counts as damaging.
In England the Environment Agency's Risk of Flooding from Rivers and Sea map is also used (2 m grid, [Open Government Licence v3.0](Open Government Licence v3.0)): for each cell the yearly likelihood of river and sea water reaching at least 0.2, 0.3, 0.6, 0.9 and 1.2 m, in four categories: high (1 in 30 or more), medium (1 in 30 to 1 in 100), low (1 in 100 to 1 in 1,000) and very low (less than 1 in 1,000). It takes into account the defences and the chance of them being overtopped or failing. Where the Environment Agency kept the previous version (50 m grid) there are no depths: there the class reads the category of any flooding (the primary map) and no 0.6 m value is shown. Wales, Scotland and Northern Ireland are not included. Contains public sector information licensed under the Open Government Licence v3.0. © Environment Agency copyright and/or database right 2025. All rights reserved.
Calculation.
(1) The water depth at the site is read in the 9 maps.
(2) The return period at which the water reaches 0.5 m is found, interpolating between two maps on the logarithm of the return period (an xNatura choice: ThinkHazard! reads only the 10, 50 and 1,000-year maps).
(3) A flood smaller than the defences does not arrive: the return period used is the larger of this one and the protection standard of the region.
(4) The class is the same in every scenario and year, because the maps are of the present climate.
(5) Where there is an official map, the first return period at which the map brings damaging water to the site's point is read. The class is the higher of the Aqueduct one and the map one: the map can only raise it. The official maps already take the existing defences into account, each in its own way (see Data), so FLOPROS is not applied.
In Switzerland a flood of medium or strong intensity counts: at least 0.5 m of water, or velocity times depth of at least 0.5 m²/s (Swiss intensity classes, Loat and Petrascheck 1997), the same 0.5 m threshold as ThinkHazard!. Map scenarios: from 5 to 300 years; the extreme scenario counts as 1,000 years. Valais publishes all water processes together (debris floods and bank erosion included) and they are used as published. Almost every canton publishes only the 30, 100 and 300-year scenarios (Ticino also 5, 10, 20, 50 and 500): there 30 means “30 years or more often”, and the map alone reaches class C at most. The map is read on a 5 m grid where a flood area counts in every cell it touches, because torrent strips are often narrower than 5 m (an xNatura choice).
In Italy P3 counts as 20 years and P2 as 100, the lower bounds of Legislative Decree 49/2010 (20 to 50 and 100 to 200 years). For P1 the decree gives no return period (“low probability of floods or extreme event scenarios”) and the districts use from just over 200 to 500 years (over 200 for the Arno and the Tuscan basins, 300 and 500 in the Po district, 300 in the Eastern Alps); 300 years are used (an xNatura choice): any value between 50 and 1,000 years gives the same class B. For P3 and P2 too some districts use return periods other than the decree's (P3 from 2 to 10 years in some units of the Po district: Po, Reno, Romagna basins; P2 from 30 years in Tuscany): there the Italian map can give a lower class than the plan. On its own the Italian map reaches class D at most.
In Austria each scenario counts with its own return period: HQ30 30 years, HQ100 100, HQ300 300; on its own the Austrian map reaches class C at most.
In England the class reads the category of water reaching at least 0.3 m, the published depth just below the 0.5 m of Aqueduct and ThinkHazard!: so the map never says a site is safer than the 0.5 m threshold would (an xNatura choice). Each category counts with its longest return period: high 30 years, medium 100, low 1,000; so 30 means “30 years or more often” and the map alone reaches class C at most. The very low category does not change the class. The same reading for 0.6 m is shown next to it.
The maps of Italy, Austria and England are read on a 10 m grid where every cell touched by an area counts.
Thresholds → level A–F.
ThinkHazard! (GFDRR) river flood hazard levels: high if 0.5 m of water arrives within 10 years, medium within 50, low within 1,000, very low beyond; translated into the six classes as for gusts and snow:
• F: at most every 3.2 years
• E: 3.2–10 years
• D: 10–22.4
• C: 22.4–50
• B: 50–1,000
• A: beyond 1,000 years, or defences of at least 1,000 years.
If the maps cannot be read, the class falls back to the rainfall indices (previous version).
From physical to impact. The physical level (A–F) is combined with the production process's sensitivity through the risk matrix (6 physical levels by 5 sensitivity bands) → impact level. The euro amount of this risk is the one described in the "Economic damage" paragraph (Version 2026): the damage factor per class of Version 2025 is no longer used for this risk.
Economic damage. Besides the class, the loss in euro is estimated as a share of the asset value (book value), with the WRI Aqueduct Floods method (Ward et al. 2020).
• Water depth: WRI Aqueduct Floods v2 river flood maps, cells of about 1 km, for 9 return periods (2, 5, 10, 25, 50, 100, 250, 500 and 1,000 years), simulated without defences. They are maps of the present climate, used for every scenario and year.
• Damage: JRC global depth-damage curves (Huizinga et al. 2017) for industrial and commercial buildings, those of the site's continent (the global curve where the continent has none), multiplied by 0.6: the JRC treats 40% of a building as not damageable by water, and here the asset value is the book value. No water, no damage. Industrial or commercial use from the ISIC sections of the site's production processes (sections A to F, that is agriculture, mining, manufacturing, energy, water and waste, construction: industrial; G to U, services: commercial; a mixed site in proportion to its processes); without processes, half and half.
• Protection: floods smaller than the protection standard of the region do no damage. FLOPROS standard (Scussolini et al. 2016): design level, else policy level, else modelled level; the Netherlands 1,000 years; where there is no value, 2 years, that is no defences.
• Expected annual loss: area under the curve of the loss against the exceedance probability, only for the floods beyond the protection level and up to the 1,000-year flood. Loss at the 100-year flood: 0 if the protection is at least 100 years. Basis: published source.
The class and the euro come from the same maps and the same defences, except where an official map raises the class: the euro stays the Aqueduct one (the depths of the official maps are not used for the euro yet), so in that case the class can be high with zero loss. The variable “Source of the flood class” says which map set the class.
Sources.
• [Copernicus Climate Data Store (CDS)](Copernicus Climate Data Store (CDS))
• [Indices for monitoring changes in extremes based on daily temperature and precipitation data](Indices for monitoring changes in extremes based on daily temperature and precipitation data)
• [ETCCDI Climate Change Indices](ETCCDI Climate Change Indices)
• [Handbook of Drought and Flood Indices](Handbook of Drought and Flood Indices)
• [Guidelines on Analysis of Extremes in a Changing Climate in Support of Adaptation to Climate Change](Guidelines on Analysis of Extremes in a Changing Climate in Support of Adaptation to Climate Change)
• [ThinkHazard! methodology report (GFDRR)](ThinkHazard! methodology report (GFDRR))
• [Ward et al. 2020, Aqueduct Floods methodology](Ward et al. 2020, Aqueduct Floods methodology)
• [Huizinga et al. 2017, Global flood depth-damage functions](Huizinga et al. 2017, Global flood depth-damage functions)
• [Scussolini et al. 2016, FLOPROS: an evolving global database of flood protection standards](Scussolini et al. 2016, FLOPROS: an evolving global database of flood protection standards)
• [Cantonal hazard maps, geodienste.ch (MGDM 166.1)](Cantonal hazard maps, geodienste.ch (MGDM 166.1))
• [Loat and Petrascheck 1997, Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten](Loat and Petrascheck 1997, Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten)
• [ISPRA (2020) Pericolosità e indicatori di rischio alluvioni, piattaforma IdroGEO](ISPRA (2020) Pericolosità e indicatori di rischio alluvioni, piattaforma IdroGEO)
• [Autorità di bacino distrettuale del fiume Po (2020), Relazione metodologica delle mappe di pericolosità e rischio di alluvione](Autorità di bacino distrettuale del fiume Po (2020), Relazione metodologica delle mappe di pericolosità e rischio di alluvione)
• [Autorità di bacino distrettuale delle Alpi Orientali, Piano di gestione del rischio di alluvioni 2021-2027](Autorità di bacino distrettuale delle Alpi Orientali, Piano di gestione del rischio di alluvioni 2021-2027)
• [Autorità di bacino distrettuale dell'Appennino Settentrionale, Relazione metodologica delle mappe di pericolosità e rischio di alluvione (II ciclo)](Autorità di bacino distrettuale dell'Appennino Settentrionale, Relazione metodologica delle mappe di pericolosità e rischio di alluvione (II ciclo))
• [BMLUK, Hochwassergefahren- und Hochwasserrisikokarten (Hochwasserrichtlinie)](BMLUK, Hochwassergefahren- und Hochwasserrisikokarten (Hochwasserrichtlinie))
• [BMLUK (2014) Hochwasser-Gefahrenkarten, fachlicher Leitfaden](BMLUK (2014) Hochwasser-Gefahrenkarten, fachlicher Leitfaden)
• [Environment Agency (2025), Risk of Flooding from Rivers and Sea](Environment Agency (2025), Risk of Flooding from Rivers and Sea)
• [Open Government Licence v3.0](Open Government Licence v3.0).