FloodAdapt Q&A
Welcome to the FloodAdapt Q&A page. On December 2nd, we held a FloodAdapt launch webinar. In case you missed it, you can rewatch the webinar below. You will also find audience questions and our answers on this page.
Rewatch the webinar
You have not yet indicated whether you want to accept or reject cookies. This means that this element cannot be displayed.
Or go directly to:
Q&A
How do I contact you with follow-up questions?
You can send an email to floodadapt@deltares.nl.
Are users able to upload their own Digital Elevation Model / terrain model to the tool?
Yes, during setup of FloodAdapt for a user’s location, their own digital elevation model (DEM) can be used.
What kinds of assumptions are included on the effectiveness of the different adaptation options?
Modeling assumptions for the different adaptation options are discussed in the documentation (see for example modeling assumptions for building-level measures like elevating, floodproofing, or buyouts).
Effectiveness can be assessed by exploring the impacts with and without an option under different events or future conditions (a storyline approach), or by evaluating the risk-reduction benefits of an adaptation option. The modeling assumptions for the benefit analysis are described in the documentation.
Are you exploring additional adaptation measures to add (e.g. wetlands, river meandering)?
Definitely! Additional adaptation options for which we already have a conceptual design are:
- Coastal nature-based solutions (wetlands, mangroves)
- Urban drainage infrastructure solutions
- Room for the river measures (meandering, levee widening, levee lowering coupled with retention areas)
New developments like these require investment, so while they are in a design phase, they can only move forward when we obtain grants or direct client assignments.
Can I add my own flood scenario results from SFINCS?
You cannot directly import flood simulation results into FloodAdapt. However, you can connect your SFINCS model to FloodAdapt and run your flood scenario through FloodAdapt.
Can you add adaptation options for specific sectors like agriculture?
It depends on the specific option, and whether that option can be represented with one of the existing adaptation-option functionalities within FloodAdapt. Please reach out to us to inquire more about specific options you have in mind.
In addition to municipal-scale applications, are there examples at regional scale in the USA (e.g. regional governments or MPOs)?
In Charleston we built the FloodAdapt system at County scale, which still ran within a couple of minutes. In principle, there is nothing prohibiting users from expanding the scale of the underlying models, although at some point memory issues may arise.
Is there a specific FloodAdapt data format required to build a custom model for a location not already in your database (e.g. a city in Indonesia)?
Yes. Click here for documentation about what is needed to build a FloodAdapt system.
What is the resolution of the model?
FloodAdapt does not come with pre-loaded baseline models. Therefore baseline models can be developed to user specifications, such as desired resolution.
Note that SFINCS uses a subgrid method, meaning the model runs on a relatively coarse computational grid for efficiency, while high-resolution elevation data are used within each grid cell to describe the relationship between water elevations and flood volumes. This allows detailed and realistic flood mapping without the computational cost of running the entire model at very high resolution.
Is it possible to tailor FloodAdapt for research projects on dynamic or robust adaptation optimization (e.g. reinforcement learning–based approaches using SFINCS)?
Yes! FloodAdapt is designed modularly and different FloodAdapt backend functionality can be accessed via API, to write tailored scripts around FloodAdapt for optimization or other research objectives. Please reach out to our team to discuss training opportunities if needed.
Is FloodAdapt applicable for flooding driven by sea-level rise, storm surge, erosion, and higher tides?
It is highly applicable for all those drivers except erosion. It can handle sea level rise, storm surge, tides, rainfall, and river discharge.
Does the risk-reduction analysis focus only on net benefits, or also on equity and distributional impacts across households?
The risk reduction benefits calculate total benefits, but also distributional impacts (spatially distributed benefits) and equity-weighted benefits.
Click here for more information about how FloodAdapt calculates risk reduction benefits.
Click here for more information on the equity-weighting.
Are there plans or examples for applying FloodAdapt in Belgium (e.g. Limburg region)?
No, not currently, although a pilot on the Meuse is anticipated in 2026. If there is interest, please reach out to our team to discuss opportunities.
What are FloodAdapt’s capabilities for snowfall and snowmelt-driven flooding?
Currently the SFINCS model does not include snowfall or snowmelt. It can be indirectly included by having a larger upstream river discharge boundary condition, to represent a what-if scenario about larger than average snowmelt/river discharges.
Can you provide a working link for the FloodAdapt newsletter signup?
Click here to sign up for our newsletter, or navigate to the Deltares FloodAdapt webpage, and click the link that says “Sign up for the FloodAdapt newsletter”.
Can the model include sector-specific or broader socioeconomic impacts?
FloodAdapt can be configured to output impacts for different parts of the community based on any attribute that has been included in your FloodAdapt setup. For example, based on income, social vulnerability, housing type, etc.
For sector-specific impacts, like agriculture or tourism, this would require additional ‘rules’ or a model that relates direct impacts to secondary impacts. Further, socioeconomic impacts such as well-being loss (which incorporates income and recovery time) have been explored in a research environment but are not yet implemented in FloodAdapt.
What does 'baseline flood information' mean in FloodAdapt? Is a separate flood simulation required?
We often talk about a ‘baseline flood model’. No separate flood simulation is required. Here’s what we mean. FloodAdapt enables non-modelers to interact with flood and impact models in an intuitive way, representing different events, measures, or future conditions.
These user choices steer an automatic updating of the baseline flood model. That is a model that must be set up when FloodAdapt is set up at a new site location.
It is a model that represents the current situation at the site (existing infrastructure, existing land use, etc.). That only needs to be set up once. From there on out, users update the model via their choices of event, future conditions, or adaptation strategy in FloodAdapt .
Are there case studies in Africa, and how were damages calibrated in data-scarce contexts?
We have applied FloodAdapt in Mozambique. Damages themselves were not calibrated, but the flood model was adjusted by evaluating flood extents and depths for known events with local stakeholders
To what extent can FloodAdapt wrap or interface with flood models other than SFINCS?
FloodAdapt was designed modularly so that it can interface with other models besides SFINCS and Delft-FIAT. It has not yet been tested in practice, but by design a new flood model would only need to develop a model-adapter that essentially translates user input to model (parameter) updates.
Please reach out to our team if you are interested in exploring an interface with an additional flood model in FloodAdapt.
Can FloodAdapt be applied to Caribbean islands, and how would baseline data be handled?
Absolutely! Setting up the baseline models in FloodAdapt in the Caribbean islands can be done using largely globally-available data (or better local data if available), except for the digital elevation model. Global DEMs are currently of insufficient resolution to provide meaningful results.
Can additional factors be incorporated, such as wildfire burn scars affecting runoff?
This could be incorporated via the land use input, which affects the roughness parameter in the flood model. So at the wildfire burn scar location the roughness would be smoother than surrounding forested areas.
Are there plans for Linux compatibility?
The FloodAdapt backend will soon (early 2026) be platform independent using Docker. The frontend has not been tested on Linux.
Are there plans to integrate stormwater infrastructure into the flood modeling?
This is a high-priority improvement for which we already have initial design plans.
Do you provide uncertainty quantification for flood depths, and how is this communicated?
Not yet. Uncertainty in flood depth estimation, but also importantly the damage estimation, is an important consideration, but not yet integrated into the FloodAdapt application.
We have done research on this from a damage estimation perspective, and the communication plan is that the central estimate is provided in the maps, but when you click on a property (for damage uncertainty), you would see a box and whisker plot capturing the uncertainty in that estimate.
Are there plans to expand FloodAdapt to extreme heat modeling?
This is not currently in our near-term plans.
Can nature-based solutions or BMPs be included as mitigation measures?
Currently, FloodAdapt can include urban green infrastructure solutions in a high-level way. This is described in our documentation. For coastal NBS like mangroves and wetlands, further development is needed.
Recent R&D in SFINCS has led to the inclusion of a module which enables the rapid calculation of wave attenuation due to vegetation. To enable coastal NBS solutions in FloodAdapt, these advances would need to be integrated into the FloodAdapt workflow, tested for usability, and validated.
