Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains

Abstract In this study we construct and compare three different deep learning (DL) models for estimating daily snow water equivalent (SWE) from high‐resolution gridded meteorological fields over the Rocky Mountain region. To train the DL models, Snow Telemetry (SNOTEL) station‐based SWE observations...

Full description

Saved in:
Bibliographic Details
Main Authors: Shiheng Duan, Paul Ullrich, Mark Risser, Alan Rhoades
Format: Article
Language:English
Published: Wiley 2024-04-01
Series:Water Resources Research
Subjects:
Online Access:https://doi.org/10.1029/2023WR035009
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849687187891683328
author Shiheng Duan
Paul Ullrich
Mark Risser
Alan Rhoades
author_facet Shiheng Duan
Paul Ullrich
Mark Risser
Alan Rhoades
author_sort Shiheng Duan
collection DOAJ
description Abstract In this study we construct and compare three different deep learning (DL) models for estimating daily snow water equivalent (SWE) from high‐resolution gridded meteorological fields over the Rocky Mountain region. To train the DL models, Snow Telemetry (SNOTEL) station‐based SWE observations are used as the prediction target. All DL models produce higher median Nash‐Sutcliffe Efficiency (NSE) values than a conceptual SWE model and interpolated gridded data sets, although mean squared errors also tend to be higher. Sensitivity of the SWE prediction to the model's input variables is analyzed using an explainable artificial intelligence (XAI) method, yielding insight into the physical relationships learned by the models. This method reveals the dominant role precipitation and temperature play in snowpack dynamics. In applying our models to estimate SWE throughout the Rocky Mountains, an extrapolation problem arises since the statistical properties of SWE (e.g., annual maximum) and geographical properties of individual grid points (e.g., elevation) differ from the training data. This problem is solved by normalizing the SWE with its historical maximum value to alleviate extrapolation for all tested DL models. Our work shows that the DL models are promising tools for estimating SWE, and sufficiently capture relevant physical relationships to make them useful for spatial and temporal extrapolation of SWE values.
format Article
id doaj-art-e61f3e59009d4a7d97c577a544b28d59
institution DOAJ
issn 0043-1397
1944-7973
language English
publishDate 2024-04-01
publisher Wiley
record_format Article
series Water Resources Research
spelling doaj-art-e61f3e59009d4a7d97c577a544b28d592025-08-20T03:22:22ZengWileyWater Resources Research0043-13971944-79732024-04-01604n/an/a10.1029/2023WR035009Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky MountainsShiheng Duan0Paul Ullrich1Mark Risser2Alan Rhoades3Atmospheric Science Graduate Group University of California Davis CA USAAtmospheric Science Graduate Group University of California Davis CA USALawrence Berkeley National Laboratory Berkeley CA USALawrence Berkeley National Laboratory Berkeley CA USAAbstract In this study we construct and compare three different deep learning (DL) models for estimating daily snow water equivalent (SWE) from high‐resolution gridded meteorological fields over the Rocky Mountain region. To train the DL models, Snow Telemetry (SNOTEL) station‐based SWE observations are used as the prediction target. All DL models produce higher median Nash‐Sutcliffe Efficiency (NSE) values than a conceptual SWE model and interpolated gridded data sets, although mean squared errors also tend to be higher. Sensitivity of the SWE prediction to the model's input variables is analyzed using an explainable artificial intelligence (XAI) method, yielding insight into the physical relationships learned by the models. This method reveals the dominant role precipitation and temperature play in snowpack dynamics. In applying our models to estimate SWE throughout the Rocky Mountains, an extrapolation problem arises since the statistical properties of SWE (e.g., annual maximum) and geographical properties of individual grid points (e.g., elevation) differ from the training data. This problem is solved by normalizing the SWE with its historical maximum value to alleviate extrapolation for all tested DL models. Our work shows that the DL models are promising tools for estimating SWE, and sufficiently capture relevant physical relationships to make them useful for spatial and temporal extrapolation of SWE values.https://doi.org/10.1029/2023WR035009snow water equivalent predictiondeep learningextrapolation
spellingShingle Shiheng Duan
Paul Ullrich
Mark Risser
Alan Rhoades
Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
Water Resources Research
snow water equivalent prediction
deep learning
extrapolation
title Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
title_full Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
title_fullStr Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
title_full_unstemmed Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
title_short Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains
title_sort using temporal deep learning models to estimate daily snow water equivalent over the rocky mountains
topic snow water equivalent prediction
deep learning
extrapolation
url https://doi.org/10.1029/2023WR035009
work_keys_str_mv AT shihengduan usingtemporaldeeplearningmodelstoestimatedailysnowwaterequivalentovertherockymountains
AT paulullrich usingtemporaldeeplearningmodelstoestimatedailysnowwaterequivalentovertherockymountains
AT markrisser usingtemporaldeeplearningmodelstoestimatedailysnowwaterequivalentovertherockymountains
AT alanrhoades usingtemporaldeeplearningmodelstoestimatedailysnowwaterequivalentovertherockymountains