This project focuses on the Nash Cascade Neural Network (NCNN), a hydrologic conceptual model designed for deep learning-based differentiable parameter learning. It aims to represent hydrologic systems intuitively and is inspired by the synchronization of hydrologic processes in modeling using concepts, physics, and neural networks.
To set up the environment for running the NCNN project, use the provided environment_cpu.yml file. This file contains all the necessary dependencies. To create the environment, run the following command:
conda env create -f environment_cpu.yml
The main code for the NCNN project is located in ncn.py. To run the project, launch the Jupyter notebook nash_cascade_neural_network.ipynb
If you use this project or refer to its findings, please cite the following conference presentation:
Frame, J. M., Bindas, T., Araki, R., Rapp, J., & Deardorff, E. (2023). On the spontaneous synchronization of hydrologic processes and hydrologic modeling. American Geophysical Union Fall Meeting 2023, session NG21B-0620.
- Jonathan M. Frame (University of Alabama)