Flood Mapping & Analysis Report
Leveraging geospatial technology to assess and mitigate flood risks.
Data & Analytical Tools
The flood mapping project leverages cutting-edge geospatial datasets and analytical tools.
Sentinel-2 Imagery
At the heart of our analysis lies Sentinel-2 imagery, a cornerstone of the European Space Agency’s Copernicus program. With a spatial resolution of 10 meters and frequent revisit times, these satellite images provide an exceptional vantage point to observe temporal changes in land surface conditions. The pre-, during-, and post-flood imagery enabled us to discern the dynamic evolution of floodwaters with remarkable precision.
ESA WorldCover
Contextual understanding of the land cover types in Porto Alegre was facilitated by the ESA Land Cover dataset. By classifying urban areas, vegetation, bare soil, and other land uses, we could overlay our flood maps with detailed land use information, offering insights into which sectors of the city were most affected by the floods.
Global Surface Water (GSW) Dataset
The GSW dataset, developed by the European Commission's Joint Research Centre, provided a historical baseline of water bodies across the globe. This resource was critical in distinguishing between permanent water features and transient floodwaters, thereby enabling a more accurate delineation of the inundation area during the El Niño event.
Geemap & Python
Python served as the computational backbone of our project—we were able to efficiently process satellite data, execute complex calculations, and generate visualizations that communicate our findings in an accessible format.
Methodology
Our flood mapping approach follows a structured workflow.
Preprocessing
Data preprocessing involved geometric and radiometric corrections of the Sentinel-2 imagery to ensure that temporal comparisons were both accurate and reliable.
MNDWI Calculation
To effectively delineate water from non-water surfaces, we computed the Modified Normalized Difference Water Index (MNDWI). This index leverages the differential reflectance properties of water in the green and short-wave infrared (SWIR) bands. A carefully selected threshold was applied to the MNDWI images, isolating pixels corresponding to water bodies. This approach provided a robust mechanism to identify flooded areas with minimal confusion from other land features.
Flood Extent Mapping
Using the thresholded MNDWI images, we generated detailed maps of the flood extent. These maps traced the progression of the floodwaters, capturing the spatial spread of the inundation.
Permanent Water Masking
To refine our analysis, we overlaid the MNDWI-derived flood extent with the historical water maps from the GSW dataset. This masking process effectively removed permanent water bodies (e.g., rivers (River Guiaba) and lakes that are a constant feature in Porto Alegre) from the analysis, ensuring that our flood maps exclusively reflected the transient inundation caused by the El Niño event.
Impact Assessment
Finally, the refined flood maps were integrated with the ESA Land Cover data. This integration allowed us to assess the impact of the flood on various land use categories—distinguishing between urban, vegetative, and other land types. The analysis provided a nuanced picture of how different sectors of Porto Alegre were impacted, informing targeted strategies for flood mitigation and urban planning..
Results
Total flood extent:
Submerged Area: 69.791 km² (equivalent to 9,800 football fields).
Temporal Progression:
Peak inundation occurred May 10–20, 2024.
Land Cover Impacts:
| Land Use Type | Area (sq km) | Percentage of total flooded(%) |
|---|---|---|
| Agricultural Land | 42.573 | 61% |
| Built-Up Areas | 1.396 | 2% |
| Wetlands | 16.052 | 23% |
| Tree Cover | 4.187 | 6% |
| Permanent Water | 0.698 | 1% |
| Grassland | 4.885 | 7% |
Key Observations:
Cropland Dominance: 61% of flooded areas were agricultural, threatening regional food security.
Wetland Loss: 23% loss of wetlands, critical natural buffers against floods.
Urban Disruption: 1.396 km² of built-up zones flooded, including residential and transport infrastructure.
Before Flood
After Flood
Flood Extent
Permanent Water
Impact Assessment
Discussion
The results reveal systemic failures in balancing de velopment with ecological and hydrological realities.
Conclusion
The 2024 Porto Alegre flood was a manifestation of ignored warnings and unsustainable land use. Croplands and wetlands bore the brunt, but urban disruptions revealed deeper governance gaps. This study provides a replicable framework for integrating satellite data, hydrological indices, and land cover analytics to guide adaptive planning.