UFNS

Urban Flood Nowcasting System (UFNS v4.0.0)

SIH26085 — High-Resolution Urban Flood Nowcasting & Drainage Coupling
Ministry of Earth Sciences (MoES) / National Centre for Medium Range Weather Forecasting (NCMRWF)

---

1. Overview

UFNS is an operational, high-performance decision support platform designed for high-resolution urban flood nowcasting. It dynamically couples:

1. Real-Time Hydrometeorological Forcing: Live Doppler Weather Radar (DWR) optical flow nowcasting, NASA Earthdata (GPM IMERG + SMAP), and OpenWeatherMap OneCall API streams.
2. Coupled 1D/2D Hydrodynamic Physics: 2D Saint-Venant shallow water equations with finite-volume topographic depression storage coupled to 1D EPA-SWMM drainage network surcharging and tidal backflow boundaries.
3. Flood-Aware Evacuation Routing: Sub-millisecond graph traversal evaluating road passability based on water depth ($h \le 0.15\text{ m}$) and velocity-depth hazard thresholds ($D \times V \le 0.35\text{ m}^2/\text{s}$).
4. Interactive AMOLED Tactical GIS Interface: 60 FPS client-side rendering with sub-frame 1-minute video interpolation, native vector AMOLED basemaps, authenticated CARTO Dark/Voyager tiles, and categorized civic infrastructure monitoring.

---

2. Polyglot System Architecture

+---------------------------------------------------------------------------------------------------+
|                                 UFNS v4.0.0 HIGH-SPEED PIPELINE                                   |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|  [ LAYER 1: GO (Golang) — Real-Time Ingestion & Streaming Daemon ]                                |
|  • File: `services/go_stream/main.go`                                                             |
|  • Concurrently polls OpenWeatherMap, NASA IMERG/SMAP, and IMD DWR Doppler Radar mosaics.         |
|  • Computes Gunnar-Farneback optical flow advection vectors in Go.                                |
|  • High-throughput WebSocket/HTTP server broadcasting telemetry with sub-millisecond scheduling.  |
|                                                                                                   |
|  [ LAYER 2: C++20 / OpenMP — Ultra-Fast Physics & Hydrodynamic Solver Core ]                      |
|  • Files: `cpp_core/solver_2d.cpp`, `cpp_core/routing.cpp`, `cpp_core/physics_engine.cpp`           |
|  • Multi-threaded OpenMP finite-volume 2D shallow water PDE solver.                                |
|  • Evaluates 1.22M-cell depth matrix in <54 ms (>22.5M cells/second throughput).                 |
|  • Sub-millisecond A* emergency evacuation routing engine with D x V passability.                 |
|  • Exported via standard C-ABI (`libufns_physics.dll` / `.so`).                                   |
|                                                                                                   |
|  [ LAYER 3: PYTHON / FASTAPI — Model Orchestration & REST Gateway ]                              |
|  • Files: `apps/api/app.py`, `apps/api/impacts.py`, `services/physics_bridge.py`                 |
|  • Zero-copy C-ABI pointer bridge (`numpy.ctypeslib`) passing raster buffers directly.            |
|  • Handles Nelder-Mead inverse calibration, CAP v1.2 XML emergency alerts, and REST APIs.         |
|                                                                                                   |
|  [ LAYER 4: REACT 18 / TYPESCRIPT / CANVAS — Tactical GIS Dashboard ]                             |
|  • Files: `apps/web/src/App.tsx`, `apps/web/src/components/MapView.tsx`                           |
|  • Infinite-resolution UTM Vector AMOLED Basemap with anti-aliased shoreline glow.                |
|  • 60 FPS smooth 1-minute video playback with linear sub-frame keyframe interpolation.            |
|  • Zero-stutter RAM cache frame pre-buffering (60 minutes horizon).                              |
+---------------------------------------------------------------------------------------------------+

---

3. Deployment Domains & Real Datasets

---

4. Key Platform Features

4.1. Ultra-High Performance Hydrodynamics Core

• C++20 SIMD Solver: Computes shallow water runoff accumulation and microtopographic depression ponding in under 54 ms for 1.22 million cells.
• SWMM 1D Pipe Flow Coupling: Simulates storm-drain conduit surcharging, manhole overflows, and coastal tide gate backpressure.

4.2. Doppler Radar & Ingestion Streaming (Go)

• Multi-Source Ingestion: Pulls live precipitation from RainViewer / IMD DWR Doppler radar, NASA GPM IMERG, and OpenWeatherMap OneCall.
• Optical Flow Nowcaster: Computes 0–30 minute storm velocity advection vectors with zero GC stutter.

4.3. Tactical AMOLED Dark GIS Interface

• Native Vector AMOLED Basemap: Renders directly in native UTM metric space ($1\text{m} = 1\text{m}$), eliminating Web Mercator distortion.
• 5 Basemap Styles: Vector AMOLED (Default), Dark Carto (Authenticated, watermark-free), Voyager (Authenticated, high-contrast street), Satellite (Esri World Imagery), and CAD Grid.
• Separated Layers: Independent continuous rainfall intensity heatmap ($0 \to 100\text{ mm/h}$) and Doppler weather radar mosaic with range rings ($3\text{km}, 6\text{km}, 12\text{km}$) and rotating azimuth beam.

4.4. Fluid Video-Like Timeline Playback

• Sub-Frame Linear Interpolation: Blends 5-min/15-min model keyframes into continuous 1-minute steps at 60 FPS.
• Zero-Stutter Pre-Buffering: Pre-fetches upcoming 60-minute forecast frames into RAM in background batches.
• Pulsing Red Sonar Loading Overlay: Real-time concentric radar telemetry during city switching and simulation execution.

---

5. Quickstart & Installation

Prerequisites

• Python: >= 3.11
• Node.js: >= 18.0
• Go (Optional for standalone streamer): >= 1.22
• C++ Compiler (Optional for native DLL build): MSVC, GCC, or Clang

1. Clone Repository & Setup Virtual Environment

git clone https://github.com/Vinamra-Mishra/UFNS-V4.git
cd UFNS-V4

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt

2. Build Frontend Dashboard

cd apps\web
npm install
npm run build
cd ..

3. Launch Operational Server

.\.venv\Scripts\uvicorn.exe apps.api.app:app --host 127.0.0.1 --port 8000

Open http://127.0.0.1:8000 in your browser.

---

6. API Reference

---

7. Repositories & Remotes

• Primary Release Repository: https://github.com/Vinamra-Mishra/UFNS-V4
• Organization Repository: https://github.com/Vynex-Labs/SIH26085-Urban-Flood-Nowcasting-System
• Release Tag: v4.0.0