Calibration of Hydrological Models + Running Climate Change Scenrios

The goal of this project was to efficently calibrate hydrological networkx models utilizing optimization techniques. I used Spotpy for the optimization alogrithms and intergrated into the pywr ecosystem.

Relevant Files

1. Script to run any climate change scenrios
2. Parent script to start calibration
3. Bulk of the calibration code with parameter setup, iterating the model, comparing results through a loss function, and tweaking parameters
4. Script that runs the model with modified parameters and formats the outputs
5. CSVs used in choosing parameters nodes and their value ranges

Results:

When I was put onto this project claibration was done by hand and took a long time to get correct. A calibrated model of Merced's model is shown below. With our new calibration approach the model's accuracy improved by 48%