Skip to content

Repository files navigation

Flood-And-Drought-Monitoring-App-GEE-Platform.

Google Earth Engine Remote Sensing GIS Satellite Data

📌 Overview

The Flood & Drought Monitoring Application is a Google Earth Engine (GEE)-based geospatial analysis platform designed to monitor, visualize, and assess the impacts of climate-induced hazards including floods and droughts. The application integrates multi-source satellite and geospatial datasets to generate:

  • Flood extent maps using Sentinel-1 Synthetic Aperture Radar (SAR)
  • Drought condition maps using rainfall anomaly analysis
  • Population exposure assessment using WorldPop datasets
  • Agricultural impact estimation using ESA WorldCover land-use data
  • Interactive hazard visualization through a Google Earth Engine user interface

The system demonstrates how cloud-based remote sensing technologies can support rapid disaster assessment, environmental monitoring, and decision-making.


🎯 Objectives

The primary objectives of this project are:

  • Develop an interactive disaster monitoring application using Google Earth Engine
  • Detect flood-affected regions using Sentinel-1 SAR imagery
  • Identify drought-prone areas using rainfall variability analysis
  • Estimate vulnerable population exposure
  • Assess potential agricultural losses
  • Provide a simple visualization dashboard for hazard interpretation

🏗️ System Architecture

            Geospatial Data Sources

Sentinel-1 SAR CHIRPS Rainfall | | | | ↓ ↓

Flood Detection     Drought Detection


     WorldPop              ESA WorldCover
        |                       |
        ↓                       ↓

Population Exposure     Cropland Impact


            |
            ↓

  Google Earth Engine Processing

            |
            ↓

     Interactive Dashboard

🛰️ Data Sources

1. Sentinel-1 SAR

Dataset: COPERNICUS/S1_GRD

Purpose:

Flood detection

Characteristics:

  • C-band Synthetic Aperture Radar
  • Cloud independent observation
  • Day/night acquisition capability
  • VV polarization analysis

Method:

Flood areas are detected based on reduced radar backscatter values:

VV Backscatter < Threshold

Potential Water/Flood Pixels


2. CHIRPS Rainfall Dataset

Dataset:

UCSB-CHG/CHIRPS/PENTAD

Purpose:

Drought monitoring through rainfall anomaly analysis.

CHIRPS provides:

  • Satellite-based rainfall estimation
  • Long-term precipitation records
  • Global coverage

Method:

Mean Annual Rainfall < Threshold

Drought Condition


3. WorldPop Population Dataset

Dataset:

WorldPop/GP/100m/pop

Purpose:

Population exposure estimation.

The system overlays hazard maps with population distribution:

Hazard Area

Population Density

Exposed Population Estimate


4. ESA WorldCover

Dataset:

ESA/WorldCover/v100

Purpose:

Agricultural impact assessment.

Cropland classification:

Land Cover Class = 40

Cropland Extraction

Flood/Drought Impact Analysis


🔬 Methodology

Flood Mapping Workflow

  1. Collect Sentinel-1 SAR imagery
  2. Filter by:
    • Study area
    • Date range
    • Instrument mode
    • Polarization
  3. Generate median radar composite
  4. Apply water threshold classification
  5. Produce flood extent layer

Drought Mapping Workflow

  1. Collect CHIRPS precipitation data
  2. Calculate rainfall statistics
  3. Apply rainfall anomaly threshold
  4. Generate drought severity map

Impact Assessment Workflow

Hazard maps are combined with:

  • Population density
  • Land cover classification

to estimate:

  • Human exposure
  • Agricultural vulnerability

🖥️ Application Features

Interactive Hazard Selection

Users can select:

  • Floods
  • Droughts
  • Cropland Loss
  • Affected Population

Visualization

The dashboard provides:

  • Hazard maps
  • Color-coded layers
  • Rainfall time-series charts
  • Regional statistics

📂 Project Structure

Flood-And-Drought-Monitoring-App-GEE-Platform/ │ ├── Flood_Drought_Monitoring_App.js │ ├── Project_Code_Structure.js │ ├── README.md │ └── LICENSE


📊 Outputs

The application generates:

Output Description
Flood Extent Map Areas potentially affected by flooding
Drought Map Low rainfall regions
Population Exposure People located in hazard zones
Cropland Loss Agricultural areas under risk
Rainfall Chart Temporal precipitation analysis

⚠️ Limitations

This prototype has several limitations:

Flood Detection

  • Fixed backscatter threshold
  • No terrain correction
  • Limited validation with ground observations

Drought Detection

  • Uses rainfall only
  • Does not include vegetation stress or soil moisture

Impact Assessment

Population exposure does not represent actual casualties or damage.

More detailed models require:

  • Flood depth estimation
  • Infrastructure data
  • Socioeconomic information

🔮 Future Improvements

Possible improvements include:

Machine Learning Integration

Add:

  • Random Forest classification
  • Deep learning segmentation
  • Flood probability prediction

Advanced Drought Indicators

Integrate:

  • NDVI anomaly
  • Vegetation Condition Index (VCI)
  • Soil Moisture Index
  • Standardized Precipitation Index (SPI)

Real-Time Monitoring

Future versions can include:

  • Weather API integration
  • Automated alerts
  • Web-based dashboards
  • Mobile notifications

📚 Related Works

1. Google Earth Engine for Disaster Monitoring

Google Earth Engine has been widely adopted for large-scale environmental monitoring because it provides cloud-based access to satellite archives and scalable geospatial computation.

Relevant applications:

  • Flood mapping
  • Drought assessment
  • Land-cover change detection
  • Climate impact analysis

2. Sentinel-1 SAR Based Flood Mapping

Many studies use Sentinel-1 SAR imagery for flood detection because radar observations are not affected by cloud cover.

Common approaches include:

  • Backscatter thresholding
  • Change detection
  • Machine learning classification

Example research areas:

  • Rapid flood mapping after extreme rainfall events
  • Near-real-time disaster response systems

3. CHIRPS-Based Drought Monitoring

CHIRPS rainfall data has been widely used for drought analysis because of its high temporal resolution and long historical coverage.

Common drought indicators:

  • Rainfall anomaly
  • SPI calculation
  • Seasonal precipitation monitoring

4. Population Exposure Assessment

Combining hazard layers with population datasets such as WorldPop is a common approach for estimating disaster vulnerability.

Applications include:

  • Flood risk assessment
  • Humanitarian planning
  • Emergency response prioritization

5. Multi-Hazard Monitoring Platforms

Recent disaster management systems combine:

  • Remote sensing
  • GIS analysis
  • Climate datasets
  • Machine learning

to create decision-support tools for governments and humanitarian organizations.


📖 References

  1. Gorelick et al. (2017).
    Google Earth Engine: Planetary-scale geospatial analysis for everyone.
    Remote Sensing of Environment.

  2. Torres et al. (2017).
    Sentinel-1 SAR mission and applications.

  3. Funk et al. (2015).
    The Climate Hazards Infrared Precipitation with Stations (CHIRPS).

  4. Tatem et al.
    WorldPop: High-resolution population datasets for development and disaster applications.

  5. Zanaga et al. (2021).
    ESA WorldCover 10 m land cover dataset.


🤝 Contribution

Contributions are welcome.

You can contribute by:

  • Improving hazard detection algorithms
  • Adding new datasets
  • Improving visualization
  • Adding validation methods

Steps:

  1. Fork the repository
  2. Create a feature branch
  3. Commit changes
  4. Submit a pull request

👨‍💻 Author

Sohel Rana

Digital Currency Investor & Technical Lead

https://github.com/SohelRana-aiub-Pro/Flood-And-Drought-Monitoring-App-GEE-Platform/

Related Resources; https://github.com/ianpdavies/cloudy_flood_prediction