0% found this document useful (0 votes)
3 views5 pages

Flood Risk Assessment Model Overview

The document outlines a conceptual risk model for flood assessment, calculating a normalized Risk Score based on hazard, exposure, and vulnerability features. It details a comprehensive algorithm and workflow involving data inputs, preprocessing, runtime pipeline steps, and decision logic for alerting users about potential flooding. The output includes risk score visualizations and 3D models to enhance situational awareness for stakeholders.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
3 views5 pages

Flood Risk Assessment Model Overview

The document outlines a conceptual risk model for flood assessment, calculating a normalized Risk Score based on hazard, exposure, and vulnerability features. It details a comprehensive algorithm and workflow involving data inputs, preprocessing, runtime pipeline steps, and decision logic for alerting users about potential flooding. The output includes risk score visualizations and 3D models to enhance situational awareness for stakeholders.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

1.

​ Final short conceptual risk model:


For each analysis unit (adaptive tile / catchment / pixel), compute a normalized Risk
Score ∈ [0,1] as a weighted combination of hazard, exposure, and vulnerability
features:

●​ Hazard: current rainfall intensity, API (antecedent), storm duration, predicted rainfall
in next window.​

●​ Propagation: flow accumulation, contributing area, slope, distance to channel/outlet.​

●​ Local susceptibility: elevation relative to local minima (depression depth), local


storage capacity (soil infiltration, imperviousness), land-use.​

●​ Boundary conditions: river/sea level, drainage capacity, tide/surge (if applicable).​

●​ Exposure: population/infrastructure presence (optional for scoring).​

●​ Adaptive thresholding: compute percentiles across AOI so scores are comparable.

2.​Full algorithm / workflow (detailed steps & pseudocode):

Inputs

●​ DEM (COG or GeoTIFF) — reprojected to local UTM meters.​

●​ Precip rasters (real-time and recent history) aligned to grid or as station observations
(Mesonet, MRMS, IMERG).​

●​ (Optional) Soil map (saturated hydraulic conductivity), land cover / imperviousness


raster, river network (NHD), tide/sea-level time series, drainage network capacity,
population/infrastructure layers.​

●​ Configuration parameters: tileing thresholds, API decay τ, weights.

Preprocessing (one-time / periodic)

1.​ Reproject DEM → UTM, ensure no-data consistent.​

2.​ Fill depressions (hydrologic fill) to remove spurious sinks (RichDEM or TauDEM).​

3.​ Compute hydrologic attributes:​


○​ slope (degrees)​

○​ flow direction (D8)​

○​ flow accumulation (pixels and area in m²)​

○​ catchment delineation / watersheds for outlets (if required)​

4.​ Compute local minima & depression depth:​

○​ For each pixel, compute elevation minus local minimum in a neighborhood /


contributing basin.​

5.​ Load land cover & soil maps, resample to DEM grid: compute
impervious_fraction, infiltration_capacity (Ks proxy).

Runtime pipeline (every ingestion step, e.g., hourly)

A. Ingest precipitation inputs

●​ Latest raster P_t and previous n rasters {P_{t-1},...,P_{t-n}} or station time


series.​

●​ If mixing raster + station: bias-correct radar/satellite using stations (simple


gauge-adjustment).​

B. Compute local rainfall features

●​ intensity = mean(P_t over unit) (mm/hr)​

●​ duration (hours of contiguous rainfall above threshold)​

●​ API (antecedant precipitation index) per unit: API_t = P_t +


exp(-Δt/τ)*API_{t-1}​

C. Routing & volume estimate

●​ For each pixel/unit compute contributing_volume = intensity *


contributing_area (m³/s proxy), incorporate rainfall duration for total volume.​

●​ flow_velocity estimate from slope (Manning or power-law): v ≈ k *


sqrt(slope) (calibrate k).​
D. Compare with storage & channel capacity

●​ available_storage = depression_volume + infiltration_capacity


* duration + drainage_capacity * duration​

●​ overflow_flag = contributing_volume > available_storage (boolean)


or compute overflow_ratio = (contributing_volume -
available_storage) / (contributing_volume + ε)​

E. Boundary conditions

●​ If within coastal buffer: include sea_level + surge and check backwater effects.​

●​ If near river gauges: include river_stage and compare to bankfull threshold.​

F. Compute normalized features (use robust percentiles 2–98):

●​ norm_intensity, norm_API, norm_rel_elev, norm_flowacc, norm_slope,


norm_impervious, norm_overflow_ratio, norm_river_stage_factor

Purpose: Make features comparable by removing units and


reducing outlier effects.

Use robust percentiles (2–98%) to scale values:​


norm_feature=(feature−p2)/(p98−p2)​

G. Risk score:​
weighted sum (see formula above). Apply hysteresis / time smoothing to reduce flicker.​
Purpose: Quantify flood/flooding hazard.

●​ Weighted sum of normalized features:​


Risk=w1⋅norm_intensity+w2⋅norm_API+...+wn⋅norm_overflowRisk = w_1 \cdot
norm\_intensity + w_2 \cdot norm\_API + ... + w_n \cdot
norm\_overflowRisk=w1​⋅norm_intensity+w2​⋅norm_API+...+wn​⋅norm_overflow
●​ Hysteresis/time smoothing: Avoids flickering risk values due to small changes
between time steps.

H. Decision logic & alerting

●​ if risk > alert_threshold and sustained > T_hours: send alert​


●​ Include severity (Low/Medium/High/Critical) by thresholds on risk percentiles.​
Purpose: Trigger warnings for users.
●​ Alert condition: If risk exceeds threshold and persists for more than
ThoursT_{hours}Thours​.​

●​ Severity levels: Based on risk percentiles (Low / Medium / High / Critical).​

●​ Outputs: Messages to users, logs, polygons affected, expected lead time.


●​

●​ ​

I. Output

●​ Raster GeoTIFF (COG) with risk score & component layers​

●​ Tiles (PNG) for visualization​

●​ Alerts + logs with affected polygons and lead time


●​ 3D Digital Twin / Model Output:

Purpose: Provide a realistic 3D simulation of terrain, water depth, and flow


dynamics.​

Data used:​

○​ DEM (Digital Elevation Model) for terrain​

○​ Rasterized water depth or overflow layers​

○​ Flow velocity and contributing volume data​

Output formats: 3D mesh (e.g., OBJ, GLTF), or interactive WebGL visualization.​

Benefits:​

○​ Enables stakeholders to “see” floods in 3D, improving situational awareness.​

○​ Can be animated over time to visualize the evolution of water levels.​


○​ Supports integration with smart city platforms or virtual training simulations.

You might also like