The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US

Study region: This study focuses on five watersheds in the southwestern United States, where April–July (AMJJ) water supply forecasts (WSFs) inform water management. Climate change has altered long-relied-upon relationships between April 1st snow water equivalent (SWE) and AMJJ water supply, threate...

Full description

Saved in:
Bibliographic Details
Main Authors: Madeline R. Pernat, Joseph Kasprzyk, Edith Zagona, Sydney D. Walker, Ben Livneh
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Journal of Hydrology: Regional Studies
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2214581825003738
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849421393411702784
author Madeline R. Pernat
Joseph Kasprzyk
Edith Zagona
Sydney D. Walker
Ben Livneh
author_facet Madeline R. Pernat
Joseph Kasprzyk
Edith Zagona
Sydney D. Walker
Ben Livneh
author_sort Madeline R. Pernat
collection DOAJ
description Study region: This study focuses on five watersheds in the southwestern United States, where April–July (AMJJ) water supply forecasts (WSFs) inform water management. Climate change has altered long-relied-upon relationships between April 1st snow water equivalent (SWE) and AMJJ water supply, threatening the skill of traditional forecasting approaches. Study focus: This work evaluates how the interaction between model type (e.g., multiple linear regression, random forest) and feature selection influences AMJJ WSF skill. Five machine learning model types are applied in each basin. A new wrapper-based feature selection method identifies the Best Feature Set—selected from a broad pool of station-based, meteorological, and climatological features—for each basin–model type combination. Results show that the most important features vary by both basin and model type, and that model types perform similarly when each is trained on its respective Best Feature Set. New hydrologic insights: April 1st SWE is most important in highly snow-dominated basins, while April 1st precipitation accumulation becomes more important in less snow-dominated systems. Station-based features from multiple lag times are consistently selected, suggesting that earlier observations provide additional predictive value. Among meteorological and climatological features, specific humidity and the Atlantic Multidecadal Oscillation are frequently selected across basins and model types, indicating broad predictive utility. Overall, results suggest that feature selection has a greater influence on forecast skill than model type choice.
format Article
id doaj-art-0bd561e3724943eebdaa7b2f23d9217c
institution Kabale University
issn 2214-5818
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Journal of Hydrology: Regional Studies
spelling doaj-art-0bd561e3724943eebdaa7b2f23d9217c2025-08-20T03:31:28ZengElsevierJournal of Hydrology: Regional Studies2214-58182025-08-016010254810.1016/j.ejrh.2025.102548The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern USMadeline R. Pernat0Joseph Kasprzyk1Edith Zagona2Sydney D. Walker3Ben Livneh4Department of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USA; Center for Advanced Decision Support for Water and Environmental Systems (CADSWES), University of Colorado Boulder, Boulder, Colorado, USA; Corresponding author at: Department of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USA.Department of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USA; Center for Advanced Decision Support for Water and Environmental Systems (CADSWES), University of Colorado Boulder, Boulder, Colorado, USADepartment of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USA; Center for Advanced Decision Support for Water and Environmental Systems (CADSWES), University of Colorado Boulder, Boulder, Colorado, USADepartment of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USADepartment of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, Boulder, Colorado, USA; Cooperative Institute for Research in Environmental Science (CIRES), University of Colorado Boulder, Boulder, Colorado, USAStudy region: This study focuses on five watersheds in the southwestern United States, where April–July (AMJJ) water supply forecasts (WSFs) inform water management. Climate change has altered long-relied-upon relationships between April 1st snow water equivalent (SWE) and AMJJ water supply, threatening the skill of traditional forecasting approaches. Study focus: This work evaluates how the interaction between model type (e.g., multiple linear regression, random forest) and feature selection influences AMJJ WSF skill. Five machine learning model types are applied in each basin. A new wrapper-based feature selection method identifies the Best Feature Set—selected from a broad pool of station-based, meteorological, and climatological features—for each basin–model type combination. Results show that the most important features vary by both basin and model type, and that model types perform similarly when each is trained on its respective Best Feature Set. New hydrologic insights: April 1st SWE is most important in highly snow-dominated basins, while April 1st precipitation accumulation becomes more important in less snow-dominated systems. Station-based features from multiple lag times are consistently selected, suggesting that earlier observations provide additional predictive value. Among meteorological and climatological features, specific humidity and the Atlantic Multidecadal Oscillation are frequently selected across basins and model types, indicating broad predictive utility. Overall, results suggest that feature selection has a greater influence on forecast skill than model type choice.http://www.sciencedirect.com/science/article/pii/S2214581825003738Water supply forecastingSeasonal forecastingMachine learningFeature selectionNested cross-validationClimate change
spellingShingle Madeline R. Pernat
Joseph Kasprzyk
Edith Zagona
Sydney D. Walker
Ben Livneh
The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
Journal of Hydrology: Regional Studies
Water supply forecasting
Seasonal forecasting
Machine learning
Feature selection
Nested cross-validation
Climate change
title The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
title_full The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
title_fullStr The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
title_full_unstemmed The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
title_short The relative importance of model type and input features for water supply forecasting in snow-dominated basins of the southwestern US
title_sort relative importance of model type and input features for water supply forecasting in snow dominated basins of the southwestern us
topic Water supply forecasting
Seasonal forecasting
Machine learning
Feature selection
Nested cross-validation
Climate change
url http://www.sciencedirect.com/science/article/pii/S2214581825003738
work_keys_str_mv AT madelinerpernat therelativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT josephkasprzyk therelativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT edithzagona therelativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT sydneydwalker therelativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT benlivneh therelativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT madelinerpernat relativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT josephkasprzyk relativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT edithzagona relativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT sydneydwalker relativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus
AT benlivneh relativeimportanceofmodeltypeandinputfeaturesforwatersupplyforecastinginsnowdominatedbasinsofthesouthwesternus