260 Advances in Geospatial Applications for Environmental and Engineering Geology (Posters)

Recent advancements in geospatial technologies have enhanced our ability to analyze, model, and predict geological and environmental processes with high precision. This session will explore the latest innovations in remote sensing, GIS, and data analytics applied to environmental engineering geosciences.

Wednesday, 22 October 2025
8:00 AM - 5:30 PM
HBGCC, Hall 1

Enabling Reproducible Multi-hazard Risk Modeling for a Rapidly Changing Refugee Landscape in Bangladesh

Paper number: 260-6
Program: Technical Sessions
Time: 8:00 AM – 8:00 AM

Authors

Abstract

We present a reproducible framework for dynamic multi-hazard risk assessment in the Kutuplaong Rohingya Refugee Camp (KTP), one of the world’s most densely populated and hazard-prone humanitarian settings. This research investigates hydro-meteorological risks—primarily shallow landslides and flash floods—before and after refugee settlement, with a focus on landscape changes driven by both anthropogenic and natural processes. Our study integrates reproducibility principles and FAIR (Findable, Accessible, Interoperable, Reusable) data practices to enable our research to be transferable worldwide across settings.

The modeling pipeline is built upon open-source tools. We utilize version-controlled Jupyter Notebooks and Conda environments, as well as R and QGIS, to support computational reproducibility. To ensure the open science of our research, we used the Zenodo repository to archive and publicly share all relevant data and code used in the workflow of this study so far. Our work begins with a slope unit (SU)-based landslide hazard model, improving upon prior grid-based assessments. A Generalized Additive Model (GAM) applied at the SU level outperforms conventional machine learning models, offering a flexible and interpretable framework. Concurrently, we estimate above-ground biomass (AGB) to quantify landscape degradation and recovery using Sentinel-2A imagery, NASA GEDI LiDAR, and the ESA Biomass product, implementing Random Forest, SVM, and XGBoost regressors.

We are now integrating this information with further observations into the landscape evolution model using Landlab. Ultimately, we will develop a coupled multi-hazard model and simulate short-term landscape evolution over the following centuries in this rapidly changing setting. The codebase will then undergo unit testing to verify the defined classes and functions, ensuring the developed code behaves correctly across different scenarios and is efficient in detecting & fixing bugs. Subsequently, this Landlab-based Python package will be published on PyPI.

 

Geological Society of America Abstracts with Programs. Vol. 57, No. 6

doi: 10.1130/abs/2025AM-414838

© Copyright 2025 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.