Supervised by:

Karem Chokmani (Regular Member (Co-researcher))

Co-supervised by:

Saeid Homayouni (Associate Member)

Research project description

Study of the Occurrence of Ice Jams in the Rivers of Southern Quebec in a Context of Climate Change.

Introduction

Ice jams pose a significant risk to rivers in cold-climate regions, such as Quebec, where winter conditions favor their formation. These events disrupt hydrological, ecological, and socio-economic functions, causing flooding, infrastructure damage, and considerable costs. Current prediction models, relying on historical data and qualitative assessments, struggle to capture the complexity of interactions between river morphology, flow conditions, and ice movement. This leaves communities vulnerable to sudden floods without adequate preparation, particularly in remote or less monitored areas.

Objectives

The main objective of this research is to study the occurrence of ice jams in southern Quebec rivers by integrating geomorphological and meteorological factors using deep learning and evaluating the impact of climate change on this phenomenon. This project is structured around three axes: (1) developing a model to characterize the predisposition of river sections to ice jams, (2) studying the current occurrence of ice jams using historical time series (1985 to today) on critical sites, and (3) analyzing the future impact of climate change on these sites using two emission scenarios (realistic and pessimistic RCP).

Study Sites

The study area includes about 36 rivers located in southern Quebec, where ice jams are frequent and disruptive. These rivers were selected due to their particular vulnerability. Each river is divided into 250-meter sections to analyze geomorphological factors influencing ice jam formation. This study also integrates meteorological data to refine flood risk predictions, thereby contributing to better risk management in this densely populated region.

Material and methods

The methodology is based on a three-step modeling approach. Step 1: Geomorphological analysis, where rivers are divided into 250 m sections to map factors influencing ice jams (sinuosity, confluences, bridges, etc.) using a CNN-LSTM model. Step 2: Integration of historical climate data (1985 to today) to assess whether meteorological conditions favor ice jam formation on geomorphologically predisposed sections. Step 3: Future climate impact assessment, applying the model to simulations based on two emission scenarios (RCP4.5 and RCP8.5) to analyze changes in ice jam frequency and location between 2050 and 2100.

References

1. De Munck, C. S., Maddock, I. P., & Visser, A. W. (2017). Modèle d’Indice de Prédisposition aux Embâcles de Glace. Journal de Recherche en Hydrologie. 2. Madaeni, F., Lhissou, R., Chokmani, K., Raymond, S.,