Raspy-Cal: A Genetic Algorithm-Based Open-Source Python Program for Automatic Calibration of HEC-RAS Hydraulic Models

Abstract:

Manual calibration of large hydraulic model domains is a tedious and time-consuming process, yet necessary for producing accurate cross-sectional water profile data. While automatic calibration programs exist for many models, no user-friendly and broadly reusable automatic calibration system currently exists for HEC-RAS. We developed Raspy-Cal, an automatic HEC-RAS calibration program using a genetic algorithm for optimization, implemented in Python. It includes a graphical user interface, configuration files, and an interactive command-line interface, as well as libraries readily usable by other Python programs. As a case study, we used Raspy-Cal to calibrate a hydraulic model of the Los Angeles River in California and its two major tributaries, Rio Hondo and Compton Creek, which had previously been manually calibrated. We calibrated the model against both the data previously used for manual calibration and against automatically-retrieved USGS gage data. We found that the automatic calibration system matched the accuracy of manual calibrations and did so in much less time and without manual intervention. The program, which is available for free under an open-source license, will facilitate fast and precise calibration of HEC-RAS models and serve as a basis for future software development.