Scalable deep learning for watershed model calibration
Abstract
Watershed models such as the Soil and Water Assessment Tool (SWAT) consist of high-dimensional physical and empirical parameters. These parameters often need to be estimated/calibrated through inverse modeling to produce reliable predictions on hydrological fluxes and states. Existing parameter estimation methods can be time consuming, inefficient, and computationally expensive for high-dimensional problems. In this paper, we present an accurate and robust method to calibrate the SWAT model (i.e., 20 parameters) using scalable deep learning (DL). We developed inverse models based on convolutional neural networks (CNN) to assimilate observed streamflow data and estimate the SWAT model parameters. Scalable hyperparameter tuning is performed using high-performance computing resources to identify the top 50 optimal neural network architectures. We used ensemble SWAT simulations to train, validate, and test the CNN models. We estimated the parameters of the SWAT model using observed streamflow data and assessed the impact of measurement errors on SWAT model calibration. We tested and validated the proposed scalable DL methodology on the American River Watershed, located in the Pacific Northwest-based Yakima River basin. Our results show that the CNN-based calibration is better than two popular parameter estimation methods (i.e., the generalized likelihood uncertainty estimation [GLUE] and the dynamically dimensioned search [DDS],more »
- Publication Date:
- Research Org.:
- Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC)
- OSTI Identifier:
- 1900374
- Report Number(s):
- PNNL-SA-176859
Journal ID: ISSN 2296-6463
- Grant/Contract Number:
- AC05-76RL01830
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Frontiers in Earth Science
- Additional Journal Information:
- Journal Volume: 10; Journal ID: ISSN 2296-6463
- Publisher:
- Frontiers Research Foundation
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 54 ENVIRONMENTAL SCIENCES; artificial neural networks; calibration; ensemble methods; deep learning; watershed hydrology
Citation Formats
Mudunuru, Maruti K., Son, Kyongho, Jiang, Peishi, Hammond, Glenn E., and Chen, Xingyuan. Scalable deep learning for watershed model calibration. United States: N. p., 2022.
Web. doi:10.3389/feart.2022.1026479.
Mudunuru, Maruti K., Son, Kyongho, Jiang, Peishi, Hammond, Glenn E., & Chen, Xingyuan. Scalable deep learning for watershed model calibration. United States. https://doi.org/10.3389/feart.2022.1026479
Mudunuru, Maruti K., Son, Kyongho, Jiang, Peishi, Hammond, Glenn E., and Chen, Xingyuan. Thu .
"Scalable deep learning for watershed model calibration". United States. https://doi.org/10.3389/feart.2022.1026479. https://www.osti.gov/servlets/purl/1900374.
@article{osti_1900374,
title = {Scalable deep learning for watershed model calibration},
author = {Mudunuru, Maruti K. and Son, Kyongho and Jiang, Peishi and Hammond, Glenn E. and Chen, Xingyuan},
abstractNote = {Watershed models such as the Soil and Water Assessment Tool (SWAT) consist of high-dimensional physical and empirical parameters. These parameters often need to be estimated/calibrated through inverse modeling to produce reliable predictions on hydrological fluxes and states. Existing parameter estimation methods can be time consuming, inefficient, and computationally expensive for high-dimensional problems. In this paper, we present an accurate and robust method to calibrate the SWAT model (i.e., 20 parameters) using scalable deep learning (DL). We developed inverse models based on convolutional neural networks (CNN) to assimilate observed streamflow data and estimate the SWAT model parameters. Scalable hyperparameter tuning is performed using high-performance computing resources to identify the top 50 optimal neural network architectures. We used ensemble SWAT simulations to train, validate, and test the CNN models. We estimated the parameters of the SWAT model using observed streamflow data and assessed the impact of measurement errors on SWAT model calibration. We tested and validated the proposed scalable DL methodology on the American River Watershed, located in the Pacific Northwest-based Yakima River basin. Our results show that the CNN-based calibration is better than two popular parameter estimation methods (i.e., the generalized likelihood uncertainty estimation [GLUE] and the dynamically dimensioned search [DDS], which is a global optimization algorithm). For the set of parameters that are sensitive to the observations, our proposed method yields narrower ranges than the GLUE method but broader ranges than values produced using the DDS method within the sampling range even under high relative observational errors. The SWAT model calibration performance using the CNNs, GLUE, and DDS methods are compared using R2 and a set of efficiency metrics, including Nash-Sutcliffe, logarithmic Nash-Sutcliffe, Kling-Gupta, modified Kling-Gupta, and non-parametric Kling-Gupta scores, computed on the observed and simulated watershed responses. The best CNN-based calibrated set has scores of 0.71, 0.75, 0.85, 0.85, 0.86, and 0.91. The best DDS-based calibrated set has scores of 0.62, 0.69, 0.8, 0.77, 0.79, and 0.82. The best GLUE-based calibrated set has scores of 0.56, 0.58, 0.71, 0.7, 0.71, and 0.8. The scores above show that the CNN-based calibration leads to more accurate low and high streamflow predictions than the GLUE and DDS sets. Our research demonstrates that the proposed method has high potential to improve our current practice in calibrating large-scale integrated hydrologic models.},
doi = {10.3389/feart.2022.1026479},
journal = {Frontiers in Earth Science},
number = ,
volume = 10,
place = {United States},
year = {Thu Nov 24 00:00:00 EST 2022},
month = {Thu Nov 24 00:00:00 EST 2022}
}
Free Publicly Available Full Text
Publisher's Version of Record
Other availability
Search WorldCat to find libraries that may hold this journal
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.
Works referenced in this record:
Model Evaluation Guidelines for Systematic Quantification of Accuracy in Watershed Simulations
journal, January 2007
- D. N. Moriasi,
- Transactions of the ASABE, Vol. 50, Issue 3
The Influence of Input Data Standardization Method on Prediction Accuracy of Artificial Neural Networks
journal, January 2016
- Anysz, Hubert; Zbiciak, Artur; Ibadov, Nabi
- Procedia Engineering, Vol. 153
The NHDPlus dataset, watershed subdivision and SWAT model performance
journal, August 2015
- Chiang, Li-Chi; Yuan, Yongping
- Hydrological Sciences Journal, Vol. 60, Issue 10
Inverse Problem Theory and Methods for Model Parameter Estimation
book, January 2005
- Tarantola, Albert
High-Quality Spatial Climate data sets for the United States and Beyond
journal, January 2000
- C. Daly,
- Transactions of the ASAE, Vol. 43, Issue 6
Deep learning convolutional neural network in rainfall–runoff modelling
journal, April 2020
- Van, Song Pham; Le, Hoang Minh; Thanh, Dat Vi
- Journal of Hydroinformatics, Vol. 22, Issue 3
Coupling surface flow with high-performance subsurface reactive flow and transport code PFLOTRAN
journal, March 2021
- Wu, Runjian; Chen, Xingyuan; Hammond, Glenn
- Environmental Modelling & Software, Vol. 137
Using Mutual Information for Global Sensitivity Analysis on Watershed Modeling
journal, October 2022
- Jiang, Peishi; Son, Kyongho; Mudunuru, Maruti K.
- Water Resources Research, Vol. 58, Issue 10
Sensitivity of optimized parameters in watershed models
journal, April 1978
- Mein, R. G.; Brown, B. M.
- Water Resources Research, Vol. 14, Issue 2
GLUE: 20 years on
journal, November 2013
- Beven, Keith; Binley, Andrew
- Hydrological Processes, Vol. 28, Issue 24
The Ensemble Kalman Filter: theoretical formulation and practical implementation
journal, November 2003
- Evensen, Geir
- Ocean Dynamics, Vol. 53, Issue 4
Exploring the exceptional performance of a deep learning stream temperature model and the value of streamflow data
journal, December 2020
- Rahmani, Farshid; Lawson, Kathryn; Ouyang, Wenyu
- Environmental Research Letters
Optimal use of the SCE-UA global optimization method for calibrating watershed models
journal, June 1994
- Duan, Qingyun; Sorooshian, Soroosh; Gupta, Vijai K.
- Journal of Hydrology, Vol. 158, Issue 3-4
A rational performance criterion for hydrological model
journal, November 2020
- Liu, Dedi
- Journal of Hydrology, Vol. 590
Five guidelines for selecting hydrological signatures
journal, September 2017
- McMillan, Hilary; Westerberg, Ida; Branger, Flora
- Hydrological Processes, Vol. 31, Issue 26
A parallelization framework for calibration of hydrological models
journal, May 2012
- Rouholahnejad, E.; Abbaspour, K. C.; Vejdani, M.
- Environmental Modelling & Software, Vol. 31
Deep learning hunts for signals among the noise
journal, May 2018
- Edwards, Chris
- Communications of the ACM, Vol. 61, Issue 6
Hydrologic prediction for urban watersheds with the Distributed Hydrology-Soil-Vegetation Model
journal, October 2008
- Cuo, Lan; Lettenmaier, Dennis P.; Mattheussen, Bernt V.
- Hydrological Processes, Vol. 22, Issue 21
Estimating Watershed Subsurface Permeability From Stream Discharge Data Using Deep Neural Networks
journal, February 2021
- Cromwell, Erol; Shuai, Pin; Jiang, Peishi
- Frontiers in Earth Science, Vol. 9
A comprehensive review of deep learning applications in hydrology and water resources
journal, August 2020
- Sit, Muhammed; Demiray, Bekir Z.; Xiang, Zhongrun
- Water Science and Technology, Vol. 82, Issue 12
A multirate mass transfer model to represent the interaction of multicomponent biogeochemical processes between surface water and hyporheic zones (SWAT-MRMT-R 1.0)
journal, January 2020
- Fang, Yilin; Chen, Xingyuan; Gomez Velez, Jesus
- Geoscientific Model Development, Vol. 13, Issue 8
Uncertainty in hydrological signatures for gauged and ungauged catchments
journal, March 2016
- Westerberg, Ida K.; Wagener, Thorsten; Coxon, Gemma
- Water Resources Research, Vol. 52, Issue 3
Groundwater Estimation from Major Physical Hydrology Components Using Artificial Neural Networks and Deep Learning
journal, December 2019
- Afzaal, Hassan; Farooque, Aitazaz A.; Abbas, Farhat
- Water, Vol. 12, Issue 1
Generalized likelihood uncertainty estimation (GLUE) using adaptive Markov Chain Monte Carlo sampling
journal, April 2008
- Blasone, Roberta-Serena; Vrugt, Jasper A.; Madsen, Henrik
- Advances in Water Resources, Vol. 31, Issue 4
A machine learning approach to emulation and biophysical parameter estimation with the Community Land Model, version 5
journal, January 2020
- Dagon, Katherine; Sanderson, Benjamin M.; Fisher, Rosie A.
- Advances in Statistical Climatology, Meteorology and Oceanography, Vol. 6, Issue 2
On the distribution of points in a cube and the approximate evaluation of integrals
journal, January 1967
- Sobol', I. M.
- USSR Computational Mathematics and Mathematical Physics, Vol. 7, Issue 4
Multi-variable SWAT model calibration with remotely sensed evapotranspiration and observed flow
journal, January 2017
- Franco, Ana Clara Lazzari; Bonumá, Nadia Bernardi
- RBRH, Vol. 22, Issue 0
Uncertainty in hydrological signatures
journal, September 2015
- Westerberg, I. K.; McMillan, H. K.
- Hydrology and Earth System Sciences, Vol. 19, Issue 9
SWAT model calibration over Cloud infrastructures using the BigEarth platform
conference, September 2017
- Bacu, Victor; Nandra, Constantin; Stefanut, Teodor
- 2017 13th IEEE International Conference on Intelligent Computer Communication and Processing (ICCP)
DART-PFLOTRAN: An ensemble-based data assimilation system for estimating subsurface flow and transport model parameters
journal, August 2021
- Jiang, Peishi; Chen, Xingyuan; Chen, Kewei
- Environmental Modelling & Software, Vol. 142
Linking hydrologic signatures to hydrologic processes: A review
journal, January 2020
- McMillan, Hilary
- Hydrological Processes, Vol. 34, Issue 6
TOSSH: A Toolbox for Streamflow Signatures in Hydrology
journal, April 2021
- Gnann, Sebastian J.; Coxon, Gemma; Woods, Ross A.
- Environmental Modelling & Software, Vol. 138
Watershed Modeling and its Applications: A State-of-the-Art Review
journal, April 2011
- Daniel, Edsel B.
- The Open Hydrology Journal, Vol. 5, Issue 1
Knowledge-Informed Deep Learning for Hydrological Model Calibration: An Application to Coal Creek Watershed in Colorado
posted_content, August 2022
- Jiang, Peishi; Shuai, Pin; Sun, Alexandar
- Hydrology and Earth System Sciences
Implications of water management representations for watershed hydrologic modeling in the Yakima River basin
journal, January 2019
- Qiu, Jiali; Yang, Qichun; Zhang, Xuesong
- Hydrology and Earth System Sciences, Vol. 23, Issue 1
Automatic model calibration
journal, January 2005
- Eckhardt, Klaus; Fohrer, Nicola; Frede, Hans-Georg
- Hydrological Processes, Vol. 19, Issue 3
Evaluating model performance: towards a non-parametric variant of the Kling-Gupta efficiency
journal, October 2018
- Pool, Sandra; Vis, Marc; Seibert, Jan
- Hydrological Sciences Journal, Vol. 63, Issue 13-14
Physics-Guided Machine Learning for Scientific Discovery: An Application in Simulating Lake Temperature Profiles
journal, May 2021
- Jia, Xiaowei; Willard, Jared; Karpatne, Anuj
- ACM/IMS Transactions on Data Science, Vol. 2, Issue 3
Modeling arid/semi-arid irrigated agricultural watersheds with SWAT: Applications, challenges, and solution strategies
journal, November 2020
- Samimi, Maryam; Mirchi, Ali; Moriasi, Daniel
- Journal of Hydrology, Vol. 590
SWAT: Model Use, Calibration, and Validation
journal, January 2012
- J. G. Arnold,
- Transactions of the ASABE, Vol. 55, Issue 4
The Variable Infiltration Capacity model version 5 (VIC-5): infrastructure improvements for new applications and reproducibility
journal, January 2018
- Hamman, Joseph J.; Nijssen, Bart; Bohn, Theodore J.
- Geoscientific Model Development, Vol. 11, Issue 8
Analysis of iterative ensemble smoothers for solving inverse problems
journal, March 2018
- Evensen, Geir
- Computational Geosciences, Vol. 22, Issue 3
Regionalization of hydrological modeling for predicting streamflow in ungauged catchments: A comprehensive review
journal, October 2020
- Guo, Yuhan; Zhang, Yongqiang; Zhang, Lu
- WIREs Water, Vol. 8, Issue 1
Model-Agnostic Interpretability with Shapley Values
conference, July 2019
- Messalas, Andreas; Kanellopoulos, Yiannis; Makris, Christos
- 2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)
Assessing watershed hydrological response to climate change based on signature indices
journal, April 2021
- Fatehifar, Atiyeh; Goodarzi, Mohammad Reza; Montazeri Hedesh, Seyedeh Sima
- Journal of Water and Climate Change, Vol. 12, Issue 6
Prospective Interest of Deep Learning for Hydrological Inference: J. Marçais and J.-R. de Dreuzy Groundwater xx, no. x: xx-xx
journal, July 2017
- Marçais, Jean; de Dreuzy, Jean-Raynald
- Groundwater, Vol. 55, Issue 5
Moving SWAT model calibration and uncertainty analysis to an enterprise Hadoop-based cloud
journal, October 2016
- Zhang, Dejian; Chen, Xingwei; Yao, Huaxia
- Environmental Modelling & Software, Vol. 84
Recent advances in convolutional neural networks
journal, May 2018
- Gu, Jiuxiang; Wang, Zhenhua; Kuen, Jason
- Pattern Recognition, Vol. 77
From calibration to parameter learning: Harnessing the scaling effects of big data in geoscientific modeling
journal, October 2021
- Tsai, Wen-Ping; Feng, Dapeng; Pan, Ming
- Nature Communications, Vol. 12, Issue 1
Dealing with Noise Problem in Machine Learning Data-sets: A Systematic Review
journal, January 2019
- Gupta, Shivani; Gupta, Atul
- Procedia Computer Science, Vol. 161
Deep learning to estimate permeability using geophysical data
journal, September 2022
- Mudunuru, M. K.; Cromwell, E. L. D.; Wang, H.
- Advances in Water Resources, Vol. 167
The Road to NHDPlus - Advancements in Digital Stream Networks and Associated Catchments
journal, February 2016
- Moore, Richard B.; Dewald, Thomas G.
- JAWRA Journal of the American Water Resources Association, Vol. 52, Issue 4
Parameter optimization for watershed models
journal, June 1976
- Johnston, P. R.; Pilgrim, D. H.
- Water Resources Research, Vol. 12, Issue 3
Google Colaboratory
book, January 2019
- Bisong, Ekaba
- Building Machine Learning and Deep Learning Models on Google Cloud Platform
On Interface Conditions for Flows in Coupled Free-Porous Media
journal, August 2019
- Nakshatrala, Kalyana B.; Joshaghani, Mohammad S.
- Transport in Porous Media, Vol. 130, Issue 2
A review of hydrologic signatures and their applications
journal, November 2020
- McMillan, Hilary K.
- WIREs Water, Vol. 8, Issue 1
Deep learning in neural networks: An overview
journal, January 2015
- Schmidhuber, Jürgen
- Neural Networks, Vol. 61
Approximating SWAT Model Using Artificial Neural Network and Support Vector Machine
journal, April 2009
- Zhang, Xuesong; Srinivasan, Raghavan; Van Liew, Michael
- JAWRA Journal of the American Water Resources Association, Vol. 45, Issue 2
Runoff conditions in the upper Danube basin under an ensemble of climate change scenarios
journal, March 2012
- Kling, Harald; Fuchs, Martin; Paulin, Maria
- Journal of Hydrology, Vol. 424-425
SALib: An open-source Python library for Sensitivity Analysis
journal, January 2017
- Herman, Jon; Usher, Will
- The Journal of Open Source Software, Vol. 2, Issue 9