HimSim.jl

A process-based hydrological model for Himalayan watersheds, with special focus on snow and glacier processes.

Himalayan watersheds are among the most hydrologically complex systems on Earth — snowmelt, glacier melt, monsoon rainfall, and high-altitude terrain all interact to control streamflow for hundreds of millions of people downstream. Standard hydrological models often lack the process representations needed to capture this complexity.

HimSim.jl is a process-based hydrological model developed in the high-performance Julia programming language, designed specifically for Himalayan catchments with significant snow and glacier contributions.

Key Features

  • Snow and glacier processes: Explicit representation of snowpack accumulation, melt, and glacial contributions to streamflow
  • ODE-based structure: Model equations set up using ModelingToolkit.jl, part of the Julia SciML ecosystem, enabling symbolic-numeric computation and automatic differentiation
  • High-performance: Implemented in Julia for computational efficiency, enabling long simulation periods and ensemble runs

Development

This model was developed during my time as a Post-Baccalaureate Research Assistant at the Center for Water Resource Studies (CWRS), Institute of Engineering, under Prof. Dr. Vishnu Prasad Pandey (Apr. 2023 – Aug. 2024).

The use of ModelingToolkit.jl allows the model to be:

  • Automatically differentiated (enabling gradient-based calibration)
  • Symbolically simplified before compilation
  • Interfaced with Julia’s rich ODE solver ecosystem (DifferentialEquations.jl)

This project laid the foundation for my current interest in differentiable and hybrid modeling approaches in hydrology.