Climate Resilience Research Initiative

Geospatial intelligence for
flood-vulnerable regions

An independent research platform integrating remote sensing, spatial analysis, and environmental science to understand climate vulnerability and adaptive capacity in South Asia.

Nandita Saha

NS

GlacioAware was developed by Nandita Saha as an interdisciplinary environmental intelligence initiative exploring climate resilience, flood vulnerability, and sustainability-focused technology in climate-sensitive regions.

Growing up in Bangladesh — one of the world's most climate-exposed countries — shaped an early and sustained interest in the intersection of environmental science, geospatial analysis, and community-level adaptation. This platform reflects an ongoing effort to translate satellite data and spatial statistics into knowledge that is accessible, reproducible, and relevant to the populations most affected by climate change.

Research interests span flood vulnerability assessment, remote sensing (SAR and multispectral), spatial autocorrelation analysis, and the development of composite vulnerability frameworks for South Asian delta regions. Current work centers on the IFVIF (Integrated Flood Vulnerability and Impact Framework) — a multi-dimensional model for district-level flood risk characterization in Bangladesh.

This work is informed by the belief that rigorous environmental science, built with human-centered values and a commitment to open data, is among the most meaningful contributions a researcher can make toward global climate adaptation.

Active & Completed Work

Ongoing research integrating satellite imagery, spatial statistics, and vulnerability frameworks across Bangladesh's flood-affected districts.

Core Framework

IFVIF — Integrated Flood Vulnerability and Impact Framework

A composite vulnerability model combining physical flood exposure, socioeconomic sensitivity, and adaptive capacity across Bangladesh's 64 districts. The framework draws on SAR-derived inundation extent, WorldPop population density, infrastructure access indices, and poverty indicators to produce a spatially explicit vulnerability score (0–1) for each administrative unit.

Manuscript in preparation  ·  Nandita Saha, 2024–2025  ·  GEE + Python + QGIS
Satellite Analysis

SAR Flood Inundation Mapping, Bangladesh 2015–2024

Temporal analysis of Sentinel-1 C-band SAR imagery to map flood inundation extent across nine flood seasons (2015–2024). Threshold-based water detection combined with change detection algorithms produces consistent, cloud-independent inundation maps at 10m resolution. Results are used as primary inputs to the IFVIF exposure sub-index.

Active  ·  Sentinel-1 via GEE  ·  Open workflows on GitHub
Vegetation & Land Cover

NDVI Temporal Analysis — Flood-Affected Agricultural Areas

Multi-temporal NDVI analysis using Sentinel-2 (10m) to track vegetation stress, post-flood crop recovery, and longer-term land degradation trends across chars and low-lying agricultural zones. Time-series decomposition isolates flood-induced stress from seasonal variation and identifies areas of persistent vegetation decline.

Active  ·  Sentinel-2 MSI  ·  2015–2024 decadal coverage
Spatial Statistics

Hotspot Analysis — District-Level Vulnerability Clustering

Application of global Moran's I and local Getis-Ord Gi* statistics to IFVIF vulnerability scores to identify whether risk is spatially clustered. Results confirm statistically significant (p<0.01) high-vulnerability clusters in northwestern and coastal districts, informing targeted policy and intervention prioritization.

Completed  ·  PySAL + GeoDa  ·  64 districts analyzed

Research Outputs

Schematic representations of key analytical outputs from IFVIF, NDVI, and hotspot analyses.

Flood Vulnerability Index — Bangladesh Districts (IFVIF, schematic)

Vulnerability index Low Moderate High Very high Critical NW Coast

Schematic district-level IFVIF composite score. Northwest and coastal districts exhibit critical vulnerability driven by inundation frequency, population density, and low adaptive capacity.

NDVI Temporal Trend — Flood-Affected Agricultural Zones, 2015–2024

0.70 0.60 0.50 0.40 0.30 2015 2016 2017 2018 2019 2020 2021 2022 2023–24 flood flood flood Non-flood zones Flood-affected chars NDVI

NDVI derived from Sentinel-2 (10m). Flood-affected riverine chars show a persistent declining trend (approx. −0.12 across the study period), with visible post-flood stress following major inundation years.

IFVIF Sub-Index Scores — High-Risk Districts (mean values)

0.25 0.50 0.75 Flood exposure 0.79 Socioeconomic sensitivity 0.70 Low adaptive capacity 0.66 Infrastructure deficit 0.57 Population exposure 0.52 Water insecurity 0.45 Mean sub-index scores (0–1), 23 high-risk districts

IFVIF sub-index decomposition showing relative contribution of each vulnerability dimension. Flood exposure and socioeconomic sensitivity are the dominant drivers in northwestern and coastal clusters.

Getis-Ord Gi* — Spatial Clusters of Flood Vulnerability

NW hotspot Coastal Not significant HH cluster Moderate Not significant Circle size proportional to Gi* z-score magnitude

Local Gi* statistics (p<0.01, FDR-corrected). Statistically significant high-high clusters concentrate in the northwest (Rangpur, Rajshahi divisions) and coastal districts (Barisal, Khulna divisions), confirming spatially structured vulnerability.

Local Relevance, Global Significance

Bangladesh is among the world's most compelling cases for climate adaptation research — not as a passive subject, but as a source of hard-won resilience knowledge.

  • 01

    Flood Vulnerability

    Positioned at the confluence of the Ganges, Brahmaputra, and Meghna deltas, Bangladesh experiences annual monsoon flooding affecting up to 30% of national territory. Extreme events (2017, 2019, 2022) inundated critical agricultural and urban zones.

  • 02

    Climate-Sensitive Communities

    Rural populations in floodplains and coastal chars face compounding risks: flood exposure, arsenic-contaminated groundwater, saline intrusion, and seasonal agricultural disruption — concentrated among communities with the least adaptive resources.

  • 03

    Adaptation Challenges

    Existing early warning systems, embankment infrastructure, and community flood management programs have reduced mortality but face limitations in addressing displacement, land loss, and structural poverty that climate change is intensifying.

  • 04

    Resilience Systems

    Community-led adaptation — floating gardens (baira), elevated homesteads, and cyclone-resistant construction — offer transferable models for other delta regions globally. Bangladesh's adaptation experience holds lessons beyond its borders.

Bangladesh represents an urgent and instructive case for environmental research. High exposure combined with a predominantly rural, dense population creates conditions demanding spatially precise intelligence — not five years from now, but in each monsoon season.

The research produced here aims to be usable by district planners, development organizations, and community adaptation programs — not only legible to academic audiences.

Bangladesh — Climate Profile

Annual flood-affected population~7 million
Territory at risk (sea-level rise)19% of land area
Major river systems700+
Climate displacement estimate, 205013 million
Global climate vulnerability rankTop 10

Selected Research Metrics

0.73
Flood Vulnerability Index
Mean IFVIF composite, high-risk districts (scale 0–1)
−0.12
NDVI Trend
Vegetation decline, flood-affected chars, 2015–2024
0.61
Moran's I
Spatial autocorrelation of vulnerability index (p < 0.01)
23
Hotspot Districts
Statistically significant high-risk clusters (Gi*, p < 0.05)
SAR Flood Mapping NDVI Analysis Moran's I Getis-Ord Gi* IFVIF Framework Population Exposure Overlay Google Earth Engine PySAL Principal Component Analysis Spatial Regression

Data Sources

All analyses use open, institutionally maintained datasets — reproducibility and transparency are core commitments of this research.

SAR / Microwave

Sentinel-1

ESA Copernicus C-band SAR for all-weather flood inundation mapping at 10m resolution.

Multispectral

Sentinel-2

ESA Copernicus optical imagery for NDVI, land cover classification, and vegetation change detection.

Earth Observation

NASA EarthData

MODIS surface reflectance, SRTM elevation, GRACE water storage, and GPM precipitation records.

Processing Platform

Google Earth Engine

Cloud-based geospatial processing environment for planetary-scale satellite time-series analysis.

Population

WorldPop

High-resolution gridded population data disaggregated by age and sex for exposure analysis.

Spatial Analysis

PySAL / GeoDa

Open-source spatial analysis libraries for autocorrelation, hotspot detection, and spatial regression.

Areas of Focus

  • Climate resilience and adaptive capacity
  • Flood vulnerability assessment
  • Remote sensing and Earth observation
  • Geospatial environmental intelligence
  • Human-centered sustainability communication
  • Environmental displacement and equity
  • Interdisciplinary climate systems research
  • Open-science and reproducible workflows
  • Bangladesh and South Asian environmental policy

Selected Output

  • Manuscript in Preparation
    Integrated Flood Vulnerability and Impact Framework (IFVIF): A Spatial Analysis of Bangladesh's High-Risk Districts
    Nandita Saha — 2024–2025
  • Environmental Writing
    Climate Intelligence, Data Access, and the Global South
  • Data & Code
    SAR Flood Mapping Workflows — Bangladesh 2015–2024
  • Visualization
    Flood Vulnerability Atlas: District-Level Risk Mapping, Bangladesh
    Climate visualization project — GEE + Python

Where the Research Is Headed

Near-Real-Time Flood Monitoring

Integrating SAR-based inundation detection with community alert systems for localized, timely flood early warning in high-risk districts.

Climate Adaptation Indicators

Developing longitudinal metrics that track not just exposure, but changing adaptive capacity over time — measuring communities' evolving resilience.

Regional Comparative Analysis

Extending the IFVIF framework to other South and Southeast Asian delta regions to enable systematic cross-national vulnerability comparison.

Responsible ML in Flood Prediction

Exploring interpretable machine learning for flood extent prediction and vegetation stress detection, with emphasis on open access and equity in use.

Get in Touch

Open to research collaboration, academic exchange, and development-oriented partnerships in climate resilience, remote sensing, and environmental intelligence.

Accessibility, Ethics & Open Science: GlacioAware is committed to accessible environmental communication — making geospatial research comprehensible to policymakers, practitioners, and the communities whose vulnerabilities are being studied. All methodologies prioritize transparency and reproducibility. Environmental data is never neutral; this research is conducted with explicit awareness of whose conditions are being analyzed and to whose benefit that analysis is directed.