Vendor comparison
Intermap is the incumbent. It builds and operates the Czech Insurers' Association flood maps, lists the Czech and Slovak insurer associations among the users of its Aquarius RMA platform, and is the name most buyers in this market already hold a contract with. This page lists the alternatives a tender will actually surface, what each publishes about its method, and where each is better or worse than Intermap. Bytero sells into the same market and is scored on the same terms as everyone else.
Water routed over a terrain model for a set of return periods, producing a depth at the building. The highest grade of flood output, and the one that decides deductibles and exclusions on a single risk. Intermap, Verisk, JBA and Fathom sell this.
Event sets and exceedance curves for a whole book, run on a modeling platform. Built to answer a capital question rather than a single-address one. Verisk on Touchstone, Aon Impact Forecasting on ELEMENTS, and Marsh Re, formerly Guy Carpenter.
The state and association maps are the baseline every model is validated against, and they are free or near free at the point of use. An address-level product joins them to building geometry and a valuation and returns a figure in money. Bytero Shield sits here.
The three groups answer different questions. A hydrodynamic map answers "how deep at this building at a given return period". A portfolio loss model answers "how much capital does this book need". An address-level product answers "what does this address cost me per year, and what is it worth". Most tenders mix the questions, which is why the matrix below scores every provider on the same columns regardless of group.
| Provider | Markets | Modeled per-address depth | Return periods | Pluvial cover | Defense modeling | API | Licensing model | Public documentation |
|---|---|---|---|---|---|---|---|---|
| Bytero Shield(this site)Alpue s. r. o. | see /coverage | Partialfluvial depth from the JRC and EFAS layers on state Floods Directive zones; pluvial by terrain susceptibility, 2D model in parts | those of the underlying state zones and JRC layers (JRC: 10 to 500 years) | YesHAND-style susceptibility everywhere, 2D pluvial model for part of the coverage | Partialprotection-standard screening plus the state zones; no defense-failure simulation | Yesone address in, valuation plus risk out | API subscription | Yesmethod pages and guides on this site |
| IntermapAquarius RMA; builds the ČAP flood maps | Czech and Slovak insurer associations listed; Central and Eastern Europe | Yesriver flood depth from 2D hydraulic modeling | N-year river extents; flash floods in four return periods (ČAP) | Yesflash-flood map, second generation | Yesstate flood-protection measures along about 9 000 km of watercourses as model inputs (ČAP) | YesRMA Web Services: per-address query, metered calls per month | cloud subscription; ČAP maps for member insurers, paid public report | Partialproduct pages and ČAP releases; depth grids and method paper not published |
| Veriskinland flood models on Touchstone | European territory list not published on the product page | Yesphysically based fluvial and pluvial models; 5 m granularity stated | event based: a 10,000-year stochastic catalogue rather than fixed-period maps | Yesoff-floodplain risk modeled | Yesdefense systems and the chance of defense failure | Partialthrough the Touchstone and Touchstone Re platforms | platform license; terms not published | Partialterritory brochures; method detail not on the product page |
| JBA Risk ManagementGlobal Flood Maps | every country at 30 m; 5 m for continental Europe | Yesextents and depths | multiple return periods; the list is in JBA documentation | Yessurface water alongside river and coastal | Yesdefended river areas from flood defense information | YesAPI, WMTS, JBA Vision viewer, dataset license | dataset license or online services | Yesexecutive briefing and knowledge base |
| Fathoma Swiss Re company | global | Yesabout 30 m globally, 10 m where regional data allows | 1-in-5 to 1-in-1000 years | Yesfluvial, pluvial and coastal | Yesdefended and undefended views | YesFathom API, portal, on-premise, partner platforms | modular by region or targeted area | Yespublished validation figures |
| Reinsurance-broker modelsAon Impact Forecasting; Marsh Re, formerly Guy Carpenter | Aon lists flood models for Austria, Czechia, Germany, Hungary, the Netherlands, Poland and Slovakia | Partialprobabilistic loss models; per-address depth is not the published product | event sets and exceedance curves rather than fixed-period maps | not published for the Central European models | YesAon: adaptation measures such as strengthened defenses | PartialAon ELEMENTS platform, Oasis-compatible | terms not published | PartialAon model brochure; Marsh Re publishes no model detail |
| ČAP flood mapsCzech Insurers’ Association, operated by Intermap | Czechia | Partialfour tariff zones per address; the depth behind the zone is not published | river N-year extents; flash floods in four return periods | Yesflash-flood map folded into the tariff zone | Yesstate protection measures along about 9 000 km of watercourses as inputs | Nopaid service for commercial use; the platform behind it is Intermap’s Aquarius RMA | member insurers; paid public report since February 2023 | PartialČAP page and press releases |
| SVP hazard mapsSlovenský vodohospodársky podnik | Slovakia | Yesdepth and velocity on each scenario line, MIKE Flood | Q5, Q10, Q50, Q100, Q1000 | Noriver scenarios only | not published | Nopublic viewer only | public viewing; reuse license not stated | Partialportal and ministry pages |
Marks describe what each provider publishes on its own site or in a public release, not a quality judgment. "Not published" means we could not find the fact on the provider's public pages; it does not mean the capability is absent. None of this replaces asking a vendor for output over your own portfolio and a validation against a flood you have claims for.
The incumbent, and the strongest single flood option in Czechia. It builds and operates the Czech Insurers' Association maps: the current RZ2023 generation was built in 2022 and 2023 on the state terrain model with discharges from more than 200 gauging stations and 2D hydraulic modeling that yields velocity and duration as well as depth, and a sixth generation of river maps plus a second generation of flash-flood maps, validated on anonymized claims from the September 2024 floods, is being delivered to eight insurers in 2026. Its commercial platform, Aquarius RMA, is a cloud subscription with per-address web services and an API metered by calls per month, and its site lists both the Czech and the Slovak insurer associations.
It holds no property valuation of its own, so buying Intermap gives you hazard without value. The ČAP maps are a market convention rather than a license you own: the depth grids behind the tariff zones are not published, and commercial use is a paid service.
Physically based fluvial and pluvial models with explicit modeling of defense systems and the chance of defense failure, a 10,000-year stochastic catalogue, and 5 m granularity stated on the product page, delivered on the Touchstone and Touchstone Re platforms. This is a probabilistic loss model rather than a hazard map, which is what capital and reinsurance work needs.
The product page we could read does not list which European territories the inland flood models cover, so a buyer in Czechia or Slovakia should ask for the territory list and for a validation against a local event before assuming coverage.
Global flood maps at 30 m for every country and 5 m for continental Europe, with extents and depths for river, surface water and coastal flooding, defended river areas built from flood defense information, and delivery by dataset license, API, WMTS or the JBA Vision viewer. With Fathom, the most completely documented option in this group.
A map vendor, not a valuation or property-data vendor. Joining JBA depth to a building's value and to your address keys is the buyer's work, and the exact list of return periods sits in JBA's documentation rather than on the product page.
The best-documented global model: about 30 m resolution with 10 m where regional data allows, return periods from 1-in-5 to 1-in-1000 years, fluvial, pluvial and coastal layers in defended and undefended views, an API, and published validation (a critical success index of about 0.75 on extent and a mean depth deviation of 0.6 m in the comparison Fathom publishes). Swiss Re integrates it into its own platforms.
Global consistency is both the strength and the limit. A 30 m global model does not resolve the small tributaries and the local defenses that a national 1 m terrain model and a state hazard study do, which is exactly where the Czech and Slovak public maps are strong.
Aon Impact Forecasting lists flood models for Austria, Czechia, Germany, Hungary, the Netherlands, Poland and Slovakia among more than forty catastrophe models on its ELEMENTS platform, which also runs Oasis-based models, and its brochure notes that its flood solutions incorporate adaptation measures such as strengthened defenses. Marsh Re, formerly Guy Carpenter, states that it pioneered flood models in Europe.
These are portfolio loss models described for reinsurance, underwriting and exposure management. They answer how much capital a book needs, not whether one address should be written; per-address output and pluvial treatment for the Central European models are not published, and license terms are not published either.
The Czech association map is the market's shared language: four tariff zones per address, with a paid property report for commercial purposes since 1 February 2023. The Slovak SVP maps are the opposite model: public Floods Directive scenario lines for Q5, Q10, Q50, Q100 and Q1000 with depth in meters and velocity in meters per second, built with the MIKE Flood hydrodynamic model and viewable by district.
Neither is an alternative to a model on its own. The ČAP zone hides the depth it is built from; the SVP lines carry depth but stop at the river scenarios, with no pluvial layer, no API and no stated reuse license.
One address in, valuation plus risk out. The fluvial layer is built on the EU Floods Directive hazard zones the states publish, with modeled depths from the JRC river flood maps and the Copernicus EFAS layers; the pluvial layer is terrain susceptibility in the HAND style, with a 2D pluvial model for part of the coverage. The output is an expected annual loss in money, per building, next to a market value and a rebuild cost from measured building geometry.
It is not hydrodynamic simulation everywhere. Where Intermap, Verisk or JBA cover an address with a routed 2D model, their per-address depth is the better number. Which markets are live is on the coverage page, and nowhere else on this site.
Where they cover an address, their depth is the better number. Intermap runs 2D hydraulic models behind the Czech insurers' zones and publishes river flood depth and N-year extents in Aquarius RMA. Verisk sells physically based fluvial and pluvial models with explicit defense-failure modeling and states 5 m granularity. JBA publishes extents and depths at 5 m across continental Europe. Bytero's fluvial depth is inherited from the JRC and Copernicus EFAS layers and from the state Floods Directive zones, and its pluvial layer is a terrain susceptibility index outside the areas that carry a 2D pluvial model. It is not hydrodynamic simulation everywhere.
If the question is "how deep at this door at a 100-year flood", and the address is inside their coverage, buy theirs.
Intermap, Verisk, JBA and Fathom sell hazard. None of them publishes a market value or a rebuild cost for the building the hazard lands on, so an underwriter or lender still joins two vendors on an address key before the number means money. That join, not the model, is where most coverage mismatches in these projects come from. It is the gap Bytero is built for: one address in, valuation plus risk out, with the loss expressed in money rather than as a zone.
Every entry is drawn from the provider's own public website or from a public release, and re-checked on review. No pricing is published here because none of these vendors lists it publicly, and nothing on this page comes from private offers. Corrections are welcome and will be applied: a vendor who thinks an entry misrepresents them can write to us and we will fix it.
