Remote Sensing & Flood Detection – HydroLAB
HydroLAB · Flood Mapping

Remote Sensing & Flood Detection

Integrating SAR radar, multispectral satellite imagery, and terrain-derived indicators for accurate, large-scale flood monitoring and surface water mapping.

Publications Downloads
SAR
Sentinel-1 & COSMO-SkyMed
Optical
Sentinel-2 · Landsat · MODIS
Fusion
Multi-source data integration
Overview

Satellite-Based Flood Monitoring

Satellite remote sensing provides synoptic, repeatable, and spatially continuous information that is essential for detecting and monitoring flood events. Unlike ground-based observations, Earth observation platforms can cover vast areas within hours of an event, offering critical information for emergency response and post-event assessment.

This research line integrates Synthetic Aperture Radar (SAR) imagery—which operates independently of weather and daylight conditions—with multispectral optical data from missions such as Sentinel-2 and Landsat. The combination of active and passive sensors addresses the fundamental trade-off between all-weather reliability (SAR) and spectral discrimination capability (optical).

A distinctive feature of this work is the fusion of remote sensing observations with geomorphic terrain descriptors, such as the Geomorphic Flood Index (GFI) and other DEM-derived features. Incorporating topographic context substantially reduces false detections and improves the spatial accuracy of flood maps, particularly in complex environments including urban and vegetated floodplains.

Key Research Approaches
📡
SAR-Based Flood Mapping
Bayesian networks and change detection algorithms applied to Sentinel-1 and COSMO-SkyMed backscatter imagery for all-weather flood delineation.
🛰️
Multispectral Water Indices
Systematic assessment of spectral water indices (NDWI, MNDWI, AWEI) across sensors and land cover types for surface water and flood extent mapping.
🔗
Multi-Source Data Fusion
Integration of SAR, optical, and geomorphic features through Random Forest and Bayesian classifiers to reduce uncertainty and improve classification robustness.
🧠
Machine Learning & Deep Learning
Deep neural networks for flood susceptibility mapping, exploiting satellite observations alongside terrain and hydrological predictors.
Methodological Workflow

From Satellite Acquisition to Flood Map

1

Image Acquisition

Retrieve SAR (Sentinel-1, COSMO-SkyMed) and/or optical (Sentinel-2, Landsat) imagery for pre- and post-event dates.

2

Pre-processing

Radiometric calibration, terrain correction (SAR), atmospheric correction (optical), co-registration of multi-temporal stacks.

3

Feature Extraction

Compute backscatter change ratios (SAR), spectral water indices (optical), and geomorphic terrain descriptors (GFI, HAND, slope).

4

Classification

Apply Bayesian networks, Random Forest, or deep learning classifiers to the multi-source feature space. Calibrate against reference flood maps.

5

Validation & Flood Map

Assess accuracy (OA, Kappa, F1-score) using independent validation data. Output binary flood extent and uncertainty maps.

Research Highlights

Key Contributions

Advancing flood detection through the synergistic use of radar, optical, and terrain information.

2016
Bayesian Network for SAR flood detection with ancillary data (IEEE TGRS)
2018
Review on advances in large-scale flood monitoring and detection
2022
Comprehensive review of multispectral water indices across sensors and land cover
2022
Deep neural network flood susceptibility mapping for Southern Italy
2024
Multi-source Random Forest with predictor robustness assessment
2026
Random Forest exploiting satellite observations and geomorphic features
Research Output

Publications

Peer-reviewed articles on remote sensing approaches for flood detection and water mapping.

2026
Mapping Flood Susceptibility Using Random Forest Exploiting Satellite Observations and Geomorphic Features
Albertini, C., Gioia, A., Iacobellis, V. & Manfreda, S.
2026 (in press)
JournalOptical + GFI
PDF 
@article{albertini2026rfgeo,
  title   = {{Mapping Flood Susceptibility Using Random Forest
              Exploiting Satellite Observations and Geomorphic Features}},
  author  = {Albertini, C. and Gioia, A. and Iacobellis, V. and
             Manfreda, Salvatore},
  year    = {2026}
}
2024
Assessing Multi-source Random Forest Classification and Robustness of Predictor Variables in Flooded Areas Mapping
Albertini, C., Gioia, A., Iacobellis, V., Petropoulos, G. P. & Manfreda, S.
Remote Sensing Applications: Society and Environment, 2024
JournalOptical + SAR
DOI 
@article{albertini2024rf,
  title   = {{Assessing Multi-source Random Forest Classification and
              Robustness of Predictor Variables in Flooded Areas Mapping}},
  author  = {Albertini, C. and Gioia, A. and Iacobellis, V. and
             Petropoulos, G. P. and Manfreda, Salvatore},
  journal = {Remote Sensing Applications: Society and Environment},
  year    = {2024},
  doi     = {10.1016/j.rsase.2024.101239}
}
2022
Detection of Surface Water and Floods with Multispectral Satellites
Albertini, C., Gioia, A., Iacobellis, V. & Manfreda, S.
Remote Sensing, 14(23), 6005, 2022
ReviewMultispectral
DOI

A comprehensive review of satellite remote sensing applications for water extent delineation and flood monitoring, focusing on freely available multispectral imagery. The performances of the most common spectral indices (NDWI, MNDWI, AWEI and others) are assessed across different land cover types to provide guidance for targeted applications. Challenges related to cloud cover, mixed pixels, and urban environments are discussed alongside opportunities offered by new-generation sensors.

@article{albertini2022surface,
  title   = {{Detection of Surface Water and Floods with
              Multispectral Satellites}},
  author  = {Albertini, C. and Gioia, A. and Iacobellis, V. and
             Manfreda, Salvatore},
  journal = {Remote Sensing},
  volume  = {14},
  number  = {23},
  pages   = {6005},
  year    = {2022},
  doi     = {10.3390/rs14236005}
}
2022
Flood Susceptibility Mapping Using a Deep Neural Network Model: The Case Study of Southern Italy
Balestra, F., Del Vecchio, M., Pirone, D., Pedone, M. A., Spina, D., Manfreda, S., Menduni, G. & Bignami, D. F.
Environmental Sciences Proceedings, 21, 36, 2022
Conference
DOI 
@inproceedings{balestra2022dnn,
  title     = {{Flood Susceptibility Mapping Using a Deep Neural
                Network Model: The Case Study of Southern Italy}},
  author    = {Balestra, F. and {Del Vecchio}, M. and Pirone, D. and
               Pedone, M. A. and Spina, D. and Manfreda, S. and
               Menduni, G. and Bignami, D. F.},
  booktitle = {Environmental Sciences Proceedings},
  volume    = {21},
  pages     = {36},
  year      = {2022},
  doi       = {10.3390/environsciproc2022021036}
}
2018
Advances in Large Scale Flood Monitoring and Detection
Manfreda, S., Samela, C., Refice, A., Tramutoli, V. & Nardi, F.
Hydrology, 5(3), 49, 2018
Review
DOI

Reviews advances in large-scale flood monitoring, covering satellite-based approaches (SAR and optical), geomorphic methods, and their integration. Discusses the complementarity between terrain-derived flood susceptibility maps and satellite-observed flood extents, and outlines perspectives for operational large-scale flood mapping systems.

@article{manfreda2018advances,
  title   = {{Advances in Large Scale Flood Monitoring and Detection}},
  author  = {Manfreda, S. and Samela, C. and Refice, A. and
             Tramutoli, V. and Nardi, F.},
  journal = {Hydrology},
  volume  = {5},
  number  = {3},
  pages   = {49},
  year    = {2018},
  doi     = {10.3390/hydrology5030049}
}
2016
A Bayesian Network for Flood Detection Combining SAR Imagery and Ancillary Data
D’Addabbo, A., Refice, A., Pasquariello, G., Lovergine, F., Capolongo, D. & Manfreda, S.
IEEE Transactions on Geoscience and Remote Sensing, 54(6), 3612–3625, 2016
JournalSAR
DOI

Proposes a Bayesian network framework that fuses SAR imagery with ancillary geospatial data—including terrain elevation, land use, and hydrological information—for flood extent detection. The probabilistic approach enables the integration of prior knowledge on flood susceptibility with SAR-derived backscatter observations, substantially reducing false positives in urban and vegetated areas.

@article{daddabbo2016bayesian,
  title   = {{A Bayesian Network for Flood Detection Combining
              SAR Imagery and Ancillary Data}},
  author  = {D'Addabbo, Annarita and Refice, Alberto and
             Pasquariello, Guido and Lovergine, Francesco P. and
             Capolongo, Domenico and Manfreda, Salvatore},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  volume  = {54},
  number  = {6},
  pages   = {3612--3625},
  year    = {2016},
  doi     = {10.1109/TGRS.2016.2598321}
}
Open Access

Downloads

PDFs freely available from the author’s repository.

PDF
Multi-source Random Forest — Flooded Areas
RSASE · 2024 · 10 MB
PDF
Surface Water & Floods — Multispectral Review
Remote Sensing · 2022 · 2 MB
PDF
Deep Neural Network — Southern Italy
Environ. Sci. Proc. · 2022
PDF
Advances in Large Scale Flood Monitoring
Hydrology · 2018
PDF
Bayesian Network — SAR Flood Detection
IEEE TGRS · 2016
Principal Investigator
SM
Salvatore Manfreda
Full Professor · University of Naples Federico II · Chair, IAHS MOXXI Working Group
ORCID Google Scholar Scopus ResearchGate HydroLAB

Attachments

About the author

He is Full Professor of Hydrology and Hydraulic Constructions at the University of Naples Federico II. He is currently chair of the IAHS MOXXI working group. His research primarily centers on hydrological modeling and monitoring. Recognizing the challenges posed by the complexity and limitations of traditional hydrological observations, he actively explores advanced and alternative monitoring techniques, such as the utilization of Unmanned Aerial Systems (UAS) coupled with image processing.