Taught Fall 2022
A hands-on introduction to major software tools and methods used across research and industry for hydrology and atmospheric science. We will take a project-based approach, learning software packages and best practices while building our own analysis for a real world problem. Major topics covered will include: Python for scientific computing, GitHub workflows and collaborative open source development, the basics of parallel and cloud computing, major data repositories and how to interact with them, and the basics of visualization. No prior coding experience is required.
Taught 2019
A set of tutorial materials on how to use the MetSim python package. MetSim is a software package that enables generation and temporal disaggregation of meteorologic forcing data meant for hydrologic models. In the tutorial I cover the gaps that MetSim covers, a brief overview of the methods employed, and then walk through several use cases for how to actually run MetSim in both command line and interactive modes. Additionally I cover some materials on how to implement MetSim for custom datasets and some interactive plotting techniques.
Written 2022
In this chapter we walk from beginning to end of how to implement so called “hybrid” models which blend machine learning with traditional modeling techniques for simulating the hydrologic cycle. Specifically, this chapter outlines the basics of automatic differentiation, it’s use in optimization and machine learning. Then we cover some basic background on numerical optimization. These pieces all come together in a proof-of-concept example for parameterizing very simple differential equations with neural networks to represent unknown relationships. We present a synthetic example describing reservoir behavior to illustrate the techniques. Following this we show how these techniques can be scaled from synthetic data to a real hydrologic model. The model that we implement has multiple soil storage components, vegetation, and is represented by a well understood set of differential equations. We use the techniques built up through the course of the chapter to practically show how to fuse machine learning techniques with more traditional techniques in a way that is physically consistent and interpretable.
This material is also available as a Jupyter Notebook
Written 2023
This is a machine learning tutorial that highlights using large-scale climate projections and recurrent neural networks to model snow. In this tutorial we will show you how to train an ML model to evaluate, anywhere in the world, the impact of climate change on snowpack. This tutorial will offer two key contributions: (1) using off-the-shelf cloud datasets to produce a model that can generate high resolution snow projections anywhere on the globe in a hindcast setting and (2) evaluating the ability to use the trained model to understand future snow conditions under multiple climate scenarios. Throughout the tutorial we will comment on design decisions which are made when deploying the trained model with the future in mind. This is primarily accomplished by defining modular components for data loading, data transformation, and model loading. As an outcome of this tutorial we hope that you will understand how to apply such methods to your own research.