CEVE 543 Fall 2025 Lab 9: HMM and NHMM Implementation

Hidden Markov Models for precipitation downscaling

1 Objectives

1. Fit Hidden Markov Model to observed precipitation data
2. Understand state assignments and parameter interpretation
3. Explore non-homogeneous HMM extensions for nonstationarity
4. Validate HMM fit and assess model performance

2 Before

ImportantInstructions

Do this before the lab date so that lab itself can go more smoothly.

3 Background and Reading

4 Tasks

Modify the code section below to address the following tasks.

1. Fit 2-state Hidden Markov Model to daily precipitation data
2. Interpret HMM states and parameters in terms of weather conditions
3. Extend to non-homogeneous HMM with time-varying transition probabilities
4. Validate model fit using statistical tests and visualizations

5 Code