Catalina Jerez

University of Colorado at Boulder;Cooperative Institute for Research in Environmental Sciences;Center for Advanced Decision Support for Water and Environmental Systems

Subject Areas: Hydrology, Water Resources, Hydroclimate Forecasting, Climate Variability, Extreme Events, Machine Learning, Water Management, Decision Support

 Recent Activity

ABSTRACT:

This HydroShare resource supports a study on seasonal-to-biennial streamflow forecasting in the Colorado River Basin. The resource contains processed forecast inputs and outputs, R functions, and example code associated with a 0–24-month lead forecasting framework for April–July naturalized flow at Lees Ferry, Arizona. The framework combines information from Ensemble Streamflow Prediction (ESP), North American Multi-Model Ensemble (NMME) forecasts, antecedent PRISM hydroclimate variables, and large-scale ocean–atmosphere climate indices.

The main experiment documented in this resource uses leave-P-year-out cross-validation with P = 1 for the 1983–2024 hindcast period. Machine-learning models, including Random Forest and Gradient Boosting Machine approaches, are evaluated using deterministic and probabilistic forecast verification metrics. The resource is intended to support reproducibility of the main forecast evaluation, including lead-dependent model performance and metric calculations. Raw external datasets are not redistributed here; users should refer to the original data providers for NMME, ESP, PRISM, and naturalized flow data.

Show More

 Contact

Resources
All 0
Collection 0
Resource 0
App Connector 0
Resource Resource
Processed Forecast Data and Evaluation Code for 0–24 Month Colorado River Streamflow Forecasts
Created: May 15, 2026, 11:24 p.m.
Authors: Jerez, Catalina · Rajagopalan, Balaji · Emerson LaJoie · Matthew Rosencrans · Sarah Baker · Miller, W. Paul · Shanahan, Seth · Edith Zagona

ABSTRACT:

This HydroShare resource supports a study on seasonal-to-biennial streamflow forecasting in the Colorado River Basin. The resource contains processed forecast inputs and outputs, R functions, and example code associated with a 0–24-month lead forecasting framework for April–July naturalized flow at Lees Ferry, Arizona. The framework combines information from Ensemble Streamflow Prediction (ESP), North American Multi-Model Ensemble (NMME) forecasts, antecedent PRISM hydroclimate variables, and large-scale ocean–atmosphere climate indices.

The main experiment documented in this resource uses leave-P-year-out cross-validation with P = 1 for the 1983–2024 hindcast period. Machine-learning models, including Random Forest and Gradient Boosting Machine approaches, are evaluated using deterministic and probabilistic forecast verification metrics. The resource is intended to support reproducibility of the main forecast evaluation, including lead-dependent model performance and metric calculations. Raw external datasets are not redistributed here; users should refer to the original data providers for NMME, ESP, PRISM, and naturalized flow data.

Show More