How 3D flood simulation works in the UAE

Why is the UAE a different flood-simulation problem?

Most off-the-shelf flood models were built for temperate, rainfall-rich catchments. The United Arab Emirates is the opposite — a hyper-arid environment where extreme rainfall events are rare but intense, drainage infrastructure was sized for normal precipitation budgets, and wadi networks behave non-linearly under sudden inflows. A model trained on European or North American watersheds produces misleading outputs when applied to Abu Dhabi, Dubai or Ras Al Khaimah.

The April 2024 event and the 14–16 March 2026 event made the cost of that mismatch visible. In March 2026, several emirates received in 48 hours the equivalent of an annual rainfall budget. Storm-drainage systems dimensioned for normal conditions were overwhelmed, road corridors became rivers, and emergency-response routing — built on assumptions of dry terrain — lost ground-truth accuracy within hours.

Gistin's methodology is built specifically for this context: hyper-arid urban hydrology, wadi non-linearity, sudden-onset extreme events, and the UAE's particular mix of dense coastal cities, inland urban corridors and fragile mangrove ecosystems. Every stage below is tuned to those constraints, not retrofitted from a temperate-climate baseline.

How does Stage 1 (Acquisition) build the unified geo-base?

Acquisition consolidates every relevant geospatial signal into a single referenced geo-base. The inputs we routinely fuse include high-resolution topography and digital elevation models, hydrological survey data, urban cadastre layers, road and rail network shapefiles, satellite imagery (including Earth-Observation sources made available through MBRSC and equivalent agencies), and live + historical weather feeds from the National Centre of Meteorology (NCM) and complementary networks.

The output is a single addressable geo-base in which every building, road segment, drainage point and ecosystem polygon carries the metadata the downstream stages require — elevation, surface type, drainage capacity, vulnerability class, and the references that allow the data to be revalidated against its source of truth.

What is the Stage 2 digital twin and how is it built?

The digital twin is the 3D structure on which simulations run. From the unified geo-base, we reconstruct buildings (including height, footprint and material class where available), road and rail networks (including surface, gradient and drainage), vegetation (including the mangrove and urban-green polygons that affect surface runoff), and coastline (including tidal datums and intertidal zones).

The fidelity threshold is set per use case. For an Abu Dhabi corridor study, sub-metre vertical accuracy on the road profile matters more than façade detail on adjacent buildings. For a coastal mangrove study, tidal datum precision matters more than urban density. The twin is structured so that locality-specific fidelity can be raised without rebuilding the base.

How does Stage 3 run 3D rainfall, flood and submersion scenarios?

Three scenario classes are run against the twin: rainfall (precipitation over time, with intensity curves), flood (surface accumulation, channel overflow, drainage saturation) and submersion (coastal, tidal or storm-surge driven). Each scenario can be driven from three weather regimes: real-time feeds for operational dispatch, historical events for calibration (April 2024 and March 2026 being our two most-used UAE calibration anchors), and forecast feeds for short-horizon anticipation.

The output is not a flat 2D flood map. It is a time-resolved 3D dataset: at each timestep we know water depth and velocity at each location in the twin, which is what enables the downstream alerting stage to fire on the specific buildings and road segments at risk — not on the catchment as a whole.

How do Stage 4 alerts target individual buildings, roads and assets?

Targeted alerts convert the simulation output into operational notifications. A municipality dispatcher does not need a flood map; they need a list of named roads to close, a list of named buildings to evacuate, and a list of rescue routes that remain passable. Thresholds are set per client — water depth on a residential ground floor, water velocity on a highway, submersion duration on a mangrove polygon — and alerts fire when a simulated value crosses the threshold for a named asset.

Because alerts are addressed to specific assets rather than to a region, the same simulation run produces different alert sets for different stakeholders: the road authority gets road-segment alerts, the housing authority gets building alerts, the environment agency gets ecosystem alerts. One pipeline, many operational outputs.

How do Stage 5 adaptation plans reshape investment decisions?

The final stage visualises inaccessible or flooded zones across the scenario set so that adaptation investment can be prioritised by impact. Which drainage upgrades resolve the most simulated alerts? Which road elevations open the most rescue routes? Which mangrove-protection projects reduce coastal-submersion risk at the highest leverage? Adaptation plans answer those questions with numbers, not opinions.

A worked example: in a typical coastal-corridor study, the model might show that a single drainage-culvert upgrade on one named arterial road resolves Stage 4 alerts across dozens of adjacent buildings and unlocks two alternative rescue routes that remained passable in the same simulated event. Without the simulation, that upgrade would compete against thirty other line items on a generic priority list; with the simulation, it carries a quantified resolved-alerts count. The numbers do the prioritising.

This is the stage where the methodology converts from a forecasting tool into a procurement-ready policy artefact. A municipality reviewing the output can carry the adaptation plan into a budget cycle as a defensible, data-grounded justification.

How does the methodology align with UAE national and emirate-level frameworks?

Each Gistin pilot is justified against at least one published UAE framework. The list of frameworks the methodology is built to advance includes the Abu Dhabi Climate Change Adaptation Plan, the UAE National Climate Change Plan 2017–2050, the UAE National Adaptation Plan 2017–2050, the UAE Net Zero 2050 Strategy, the UAE Nationally Determined Contribution (NDC), the UAE National Biodiversity Strategy, the UAE National Strategy for Coastal and Marine Environment, the Abu Dhabi Climate Change Strategy 2023–2027, the Abu Dhabi 2030 Urban Structure Framework Plan, the Dubai 2040 Urban Master Plan and the UAE's Green Agenda 2030.

The G3DAR research foundation, led at the United Arab Emirates University (UAEU), was specifically recognised by the Emirati Society of GIS and Remote Sensing (ESGRS) and the Mohammed Bin Rashid Space Centre (MBRSC) for this alignment when it won 1st place at the ESGRS Challenge — Earth Observation Edition.

Frequently asked questions