pith. sign in

arxiv: 2502.12396 · v1 · submitted 2025-02-18 · ⚛️ physics.flu-dyn · cs.CE· cs.LG

Scientific Machine Learning of Flow Resistance Using Universal Shallow Water Equations with Differentiable Programming

Xiaofeng Liu , Yalan Song This is my paper

Pith reviewed 2026-05-23 03:12 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn cs.CEcs.LG
keywords shallow water equationsManning's roughness coefficientuniversal differential equationsdifferentiable programminginverse modelingscientific machine learningflow resistancehybrid hydrodynamics modeling
0
0 comments X

The pith

Embedding neural networks inside shallow water equations via differentiable programming creates a solver that inverts Manning's roughness from data and learns generalizable flow-resistance relations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a hybrid solver called Hydrograd that treats the two-dimensional shallow water equations as the fixed physics backbone while allowing a neural network to learn the relationship between Manning's roughness coefficient n, hydraulic variables, and observed flow. This universal differential equation approach supports forward simulation, automatic differentiation for gradient computation, and inverse modeling to estimate spatially and temporally varying roughness without relying on empirical tables. A sympathetic reader would care because accurate, data-adapted roughness values improve flood forecasts and river engineering, and the retained physics backbone is claimed to avoid the generalization failures common in pure surrogate models that require heavy pretraining.

Core claim

The authors introduce Hydrograd, a universal shallow water equations solver built on the universal differential equation concept. The model performs accurate forward hydrodynamic simulations, uses automatic differentiation to obtain gradients for sensitivity analysis and inverse estimation of Manning's n from real river observations, and trains a neural network to capture a universal mapping from n and hydraulic parameters to flow. Because the two-dimensional shallow water equations remain the physics core, the learned relationship is presented as generalizable to out-of-sample conditions without the data-intensive pretraining required by surrogate-only methods.

What carries the argument

Universal shallow water equations (USWEs) formed by embedding a neural network inside the Manning friction term of the 2D shallow water equations, made differentiable through automatic differentiation for hybrid forward-inverse-scientific-ML use.

If this is right

  • The hybrid solver reproduces observed water levels and discharges with accuracy comparable to conventional SWE codes in forward runs.
  • Automatic differentiation supplies exact gradients that enable efficient sensitivity analysis of model outputs to Manning's n and other parameters.
  • Inverse modeling recovers spatially varying Manning's n fields that improve agreement with observed flow data in a real river application.
  • A neural network trained inside the solver learns a mapping from n and local hydraulics to flow that can be reused without retraining the physics component.
  • The overall framework supplies a route to inverse problems and physics discovery that keeps the governing equations intact rather than replacing them with data-driven surrogates.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same embedding technique could be applied to other uncertain terms such as turbulence closures or sediment transport coefficients to test whether additional universal relations emerge.
  • Real-time assimilation of new gauge data could continuously update the learned roughness mapping during an evolving flood event.
  • Cross-validation across multiple distinct river basins would provide a direct test of whether the learned mapping transfers beyond the single channel used in the study.
  • Coupling the differentiable solver to optimization routines could discover entirely new functional forms for resistance that replace the empirical Manning formula.

Load-bearing premise

The neural network can discover a relationship between Manning's n, hydraulic parameters, and flow that remains valid and accurate when applied to river channels or flow conditions different from the training data.

What would settle it

Train the neural network on one river reach, then apply the resulting model to an independent reach with its own measured discharges and water levels; large, systematic mismatches between simulated and observed flows would show the learned relationship is not universal.

Figures

Figures reproduced from arXiv: 2502.12396 by Xiaofeng Liu, Yalan Song.

Figure 1
Figure 1. Figure 1: The code structure of Hydrograd. 2.2 Mathematical Formulation for the Universal SWEs Solver 2.2.1 Governing Equations The depth-averaged SWEs are a set of partial differential equations that describe the flow in a channel with a free surface. A schematic diagram of the shallow flow and definitions are shown in [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The schematic diagram of the shallow flow and definitions: (a) the definitions of the variables; (b) the scheme of the universal shallow water equations (USWEs). which still maintains the hyperbolic nature of SWEs and works well with Roe’s Riemann solver to produce balanced momentum fluxes. The core of the idea is that the pressure gradient term is split into two parts: the first part is the pressure gradi… view at source ↗
Figure 3
Figure 3. Figure 3: The comparison of the forward simulation results of the 1D channel with a bump. For demonstration purpose, the whole domain is hypothetically divided into five zones with different Manning’s n values (see [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The forward simulation results of the Savannah River case: (a) Bathymetry, (b) Five zones of Manning’s n, (c) Simulated WSE, (d) Simulated flow velocity. The sensitivity of the model solution Q with respect to the Manning’s roughness coefficient n = [n1, n2, n3, n4, n5], i.e., the Jacobian matrix J = ∂Q/∂n, is computed –11– [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The sensitivity analysis of the Savannah River case. Each subplot shows the sensitivity of the flow variables (W SE, hu, hv) to the Manning’s n in different zones. Each column is for a flow variable and each row is for a roughness zone. 3.3 Parameter Inversion The Savannah River case shown above for the sensitivity analysis is also used to demonstrate parameter inversion. The constant Manning’s n values in… view at source ↗
Figure 6
Figure 6. Figure 6: The histories of losses for the Manning’s n inversion for the Savannah River case. The inversion process can be visualized by plotting the trajectories of the Manning’s n values [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The comparison of the simulated flow field from the final inversion step with the observed data for the Savannah River Case. The first column shows the truth (training data) and the second column shows the simulated flow field with the final inverted Man￾ning’s n values. The third column shows the difference between the simulated flow field and the truth [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Inversion of Manning’s n values: (a) The history of the losses for the Manning’s n inversion, (b) The trajectory of the Manning’s n during the inversion process (only the pair of n1 and n5 is shown). –14– [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The result of the one-dimensional channel with a bump using USWEs: (a) Com￾parison of inverted WSE profile with the truth, (b) Training loss history, (c) Comparison of inverted Manning’s n values with the truth, and (d) Comparison of the NN prediction of Manning’s n with the truth of the Sigmoid function in Equation 6. Manning’s n were computed for each cell in the mesh at each time step. Thus, even within… view at source ↗
Figure 10
Figure 10. Figure 10: The contour plot of the roughness height ks for the Savannah River case. tion 7. In the plots, each scatter point represents a computational cell in the domain. The lines and the scatters are colored with h/ks. The two plots are almost identical, indicat￾ing the NN in USWEs fully recovered the flow resistance law embedded in the training data. The scatter points cover a wide range of Re and h/ks within th… view at source ↗
Figure 11
Figure 11. Figure 11: The training history using USWEs for the Savannah River case: (a) The his￾tory of the losses, (b) The comparison of the Manning’s n values from the NN at the 150th iteration and the truth. (c) The distribution of the truth friction factor f for each cell in the domain on the f(Re, h/ks) plane. The lines are from Equation 7 and the scat￾ters are from all computational cells. Both the lines and the scatters… view at source ↗
Figure 12
Figure 12. Figure 12: The comparison of the velocity component u between the truth and the USWEs’ prediction at four different iterations during the training process. The first column shows the truth, the second column shows the USWEs’ prediction, and the third column shows the difference between the two [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: The comparison of the Manning’s n values between the truth and the USWEs’ prediction at four different iterations during the training process. The first column shows the truth, the second column shows the USWEs’ prediction, and the third column shows the difference between the two. –19– [PITH_FULL_IMAGE:figures/full_fig_p019_13.png] view at source ↗
read the original abstract

Shallow water equations (SWEs) are the backbone of most hydrodynamics models for flood prediction, river engineering, and many other water resources applications. The estimation of flow resistance, i.e., the Manning's roughness coefficient $n$, is crucial for ensuring model accuracy, and has been previously determined using empirical formulas or tables. To better account for temporal and spatial variability in channel roughness, inverse modeling of $n$ using observed flow data is more reliable and adaptable; however, it is challenging when using traditional SWE solvers. Based on the concept of universal differential equation (UDE), which combines physics-based differential equations with neural networks (NNs), we developed a universal SWEs (USWEs) solver, Hydrograd, for hybrid hydrodynamics modeling. It can do accurate forward simulations, support automatic differentiation (AD) for gradient-based sensitivity analysis and parameter inversion, and perform scientific machine learning for physics discovery. In this work, we first validated the accuracy of its forward modeling, then applied a real-world case to demonstrate the ability of USWEs to capture model sensitivity (gradients) and perform inverse modeling of Manning's $n$. Furthermore, we used a NN to learn a universal relationship between $n$, hydraulic parameters, and flow in a real river channel. Unlike inverse modeling using surrogate models, Hydrograd uses a two-dimensional SWEs solver as its physics backbone, which eliminates the need for data-intensive pretraining and resolves the generalization problem when applied to out-of-sample scenarios. This differentiable modeling approach, with seamless integration with NNs, provides a new pathway for solving complex inverse problems and discovering new physics in hydrodynamics.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces Hydrograd, a universal shallow water equations (USWEs) solver based on the universal differential equations (UDE) framework. It integrates a 2D physics-based SWE solver with neural networks to enable forward simulations, automatic differentiation for sensitivity analysis and parameter inversion, and scientific machine learning to discover a relationship for Manning's roughness coefficient n from hydraulic parameters and flow data. The central claim is that this physics-backbone approach eliminates data-intensive pretraining required by surrogate models and resolves out-of-sample generalization for the learned n-mapping, as demonstrated via validation of forward modeling and a real-world river channel application.

Significance. If the generalization result holds, the work provides a technically promising hybrid modeling pathway for hydrodynamics that preserves physical consistency via the differentiable SWE backbone while allowing data-driven components for roughness. This could advance inverse problems in flood prediction and river engineering. The seamless AD integration for gradients is a clear strength of the differentiable programming approach.

major comments (2)
  1. [Abstract] Abstract: The claim that Hydrograd 'resolves the generalization problem when applied to out-of-sample scenarios' is load-bearing for the central contribution yet unsupported; the manuscript reports NN training only on data from one real river channel, with no quantitative out-of-sample results (e.g., error metrics or comparisons) on a second channel, altered geometry, or different flow regime provided to test the universality of the learned n-relationship.
  2. [Abstract] Abstract: Validation of forward modeling accuracy and real-world inverse modeling of n is asserted, but no quantitative error metrics, baseline comparisons (e.g., against empirical n formulas or traditional solvers), or details on how NN integration affects SWE accuracy and stability are supplied, preventing verification of the hybrid model's performance claims.
minor comments (1)
  1. [Abstract] The abstract would be strengthened by including at least one key quantitative result (e.g., forward-model error or inversion accuracy) to allow readers to assess the claims without the full text.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment below and will revise the manuscript to improve clarity and accuracy in the abstract while preserving the core contributions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that Hydrograd 'resolves the generalization problem when applied to out-of-sample scenarios' is load-bearing for the central contribution yet unsupported; the manuscript reports NN training only on data from one real river channel, with no quantitative out-of-sample results (e.g., error metrics or comparisons) on a second channel, altered geometry, or different flow regime provided to test the universality of the learned n-relationship.

    Authors: We agree that the manuscript does not provide quantitative out-of-sample testing on a second channel or altered conditions, so the strong claim in the abstract is not fully supported by additional empirical evidence. The argument in the paper is that embedding the NN inside the physics-based 2D SWE solver (rather than training a standalone surrogate) removes the need for data-intensive pretraining and thereby mitigates typical generalization failures of pure ML models; this is demonstrated on one real channel. To avoid overstatement we will revise the abstract to qualify the statement, emphasizing the architectural advantage for generalization while noting that multi-channel validation remains future work. A limitations paragraph will also be added. revision: yes

  2. Referee: [Abstract] Abstract: Validation of forward modeling accuracy and real-world inverse modeling of n is asserted, but no quantitative error metrics, baseline comparisons (e.g., against empirical n formulas or traditional solvers), or details on how NN integration affects SWE accuracy and stability are supplied, preventing verification of the hybrid model's performance claims.

    Authors: The full manuscript contains quantitative forward-modeling error metrics (L2 norms versus analytical and traditional solvers) and inverse-modeling results in Sections 3 and 4. However, these numbers are not summarized in the abstract. We will revise the abstract to include the key reported accuracy figures and a brief statement on stability. We will also add an explicit sentence clarifying that NN integration does not degrade the underlying SWE solver's accuracy or stability beyond the levels already shown in the validation experiments. revision: yes

Circularity Check

0 steps flagged

No circularity; physics backbone and NN fit remain independent

full rationale

The derivation chain rests on standard shallow water equations as an external physics component combined with a neural network trained on observed flow data to infer Manning's n. Forward simulation accuracy, gradient computation, and inverse modeling are validated against the SWE solver itself, which is not constructed from the NN outputs. The generalization claim is asserted as a benefit of retaining the full physics solver rather than surrogates, but this does not reduce any prediction to the training fit by definition or via self-citation. No load-bearing self-citations, ansatzes, or renamings appear in the provided text.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the standard validity of 2D shallow water equations for the target flows and on the capacity of a neural network to extract a generalizable mapping for n without the paper providing independent verification of that mapping.

free parameters (1)
  • Neural network weights and biases
    Learned from observed flow data to represent the relationship for Manning's n.
axioms (1)
  • domain assumption The 2D shallow water equations provide an accurate physics backbone for the river channel flows considered.
    Invoked as the fixed component that the NN augments.

pith-pipeline@v0.9.0 · 5835 in / 1229 out tokens · 30009 ms · 2026-05-23T03:12:37.402354+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

68 extracted references · 68 canonical work pages · 1 internal anchor

  1. [1]

    AboelyazeedEtAl2023 APACrefauthors Aboelyazeed, D. , Xu, C. , Hoffman, F M. , Liu, J. , Jones, A W. , Rackauckas, C. Shen, C. APACrefauthors \ 2023 . A differentiable, physics-informed ecosystem modeling and learning framework for large-scale inverse problems: demonstration with photosynthesis simulations A differentiable, physics-informed ecosystem model...

    work page doi:10.5194/bg-20-2671-2023 2023

  2. [2]

    , Borchers, B

    Aster2013 APACrefauthors Aster, R C. , Borchers, B. \ Thurber, C H. APACrefauthors \ 2013 . Parameter Estimation and Inverse Problems Parameter estimation and inverse problems \ ( Second \ ). Elsevier Inc

    work page 2013

  3. [3]

    , Andres, N

    badoux2014damage APACrefauthors Badoux, A. , Andres, N. \ Turowski, J. APACrefauthors \ 2014 . Damage costs due to bedload transport processes in Switzerland Damage costs due to bedload transport processes in switzerland . Natural Hazards and Earth System Sciences 14 2 279--294

    work page 2014

  4. [4]

    APACrefauthors \ 1977

    Barr1977 APACrefauthors Barr, D I H. APACrefauthors \ 1977 . Discussion of ``Accurate explicit equation for friction factor'' Discussion of ``accurate explicit equation for friction factor'' . Journal of the Hydraulics Division 103 HY3 334--337

    work page 1977

  5. [5]

    APACrefauthors \ 1985

    bathurst1985flow APACrefauthors Bathurst, J C. APACrefauthors \ 1985 . Flow resistance estimation in mountain rivers Flow resistance estimation in mountain rivers . Journal of Hydraulic Engineering 111 4 625--643

    work page 1985

  6. [6]

    , Tsai, W P

    bindas2024improving APACrefauthors Bindas, T. , Tsai, W P. , Liu, J. , Rahmani, F. , Feng, D. , Bian, Y. Shen, C. APACrefauthors \ 2024 . Improving river routing using a differentiable Muskingum-Cunge model and physics-informed machine learning Improving river routing using a differentiable muskingum-cunge model and physics-informed machine learning . Wat...

    work page 2024

  7. [7]

    APACrefauthors \ 1995

    brunner1995hec APACrefauthors Brunner, G W. APACrefauthors \ 1995 . HEC-RAS River Analysis System: Hydraulic Reference Manual, Version 1.0 HEC-RAS river analysis system: Hydraulic reference manual, version 1.0 \ Technical Report . Davis, CA U.S. Army Corps of Engineers, Hydrologic Engineering Center

    work page 1995

  8. [8]

    , Graham, L

    butler2015definition APACrefauthors Butler, T. , Graham, L. , Estep, D. , Dawson, C. \ Westerink, J. APACrefauthors \ 2015 . Definition and solution of a stochastic inverse problem for the Manning’sn parameter field in hydrodynamic models Definition and solution of a stochastic inverse problem for the manning’sn parameter field in hydrodynamic models . Ad...

    work page 2015

  9. [9]

    , Goswami, S

    cao2024laplace APACrefauthors Cao, Q. , Goswami, S. \ Karniadakis, G E. APACrefauthors \ 2024 . Laplace neural operator for solving differential equations Laplace neural operator for solving differential equations . Nature Machine Intelligence 6 6 631--640

    work page 2024

  10. [10]

    , Yulia, R

    chen2018neural APACrefauthors Chen, T. , Yulia, R. , Jesse, B. \ K, D D. APACrefauthors \ 2018 . Neural ordinary differential equations Neural ordinary differential equations . Advances in neural information processing systems 31

    work page 2018

  11. [11]

    , DiBiase, R

    ChenEtAl2019 APACrefauthors Chen, Y. , DiBiase, R. , McCarroll, N. \ Liu, X. APACrefauthors \ 2019 . Quantifying flow resistance in mountain streams using computational fluid dynamics modeling over structure-from-motion photogrammetry derived microtopography Quantifying flow resistance in mountain streams using computational fluid dynamics modeling over s...

    work page doi:10.1002/esp.4624 2019

  12. [12]

    APACrefauthors \ 2008

    Cheng2008 APACrefauthors Cheng, N S. APACrefauthors \ 2008 . Formulas for friction factor in transitional regimes Formulas for friction factor in transitional regimes . Journal of Hydraulic Engineering 134 9 1357--1362

    work page 2008

  13. [13]

    APACrefauthors \ 1775

    Chezy1775 APACrefauthors Ch\' e zy, A. APACrefauthors \ 1775 . Formula to find the uniform velocity that the water will have in a ditch or in a canal of which the slope is known Formula to find the uniform velocity that the water will have in a ditch or in a canal of which the slope is known . Paris, France . Manuscript 847, \' E cole des P onts et C haus...

    work page 1921

  14. [14]

    APACrefauthors \ 1973

    Churchill1973 APACrefauthors Churchill, S W. APACrefauthors \ 1973 . Empirical expressions for the shear stress in turbulent flow in commercial pipe Empirical expressions for the shear stress in turbulent flow in commercial pipe . AIChE Journal 19 2 375--376 . APACrefDOI doi:10.1002/aic.690190228 APACrefDOI

    work page doi:10.1002/aic.690190228 1973

  15. [15]

    \ White, C M

    ColebrookWhite1937 APACrefauthors Colebrook, C F. \ White, C M. APACrefauthors \ 1937 . Experiments with fluid friction in roughened pipes Experiments with fluid friction in roughened pipes . Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences 161 367--381 . APACrefDOI doi:10.1098/rspa.1937.0150 APACrefDOI

    work page doi:10.1098/rspa.1937.0150 1937

  16. [16]

    \ Rackauckas, C

    optimization2023 APACrefauthors Dixit, V K. \ Rackauckas, C. APACrefauthors \ 2023 mar . Optimization.jl: A Unified Optimization Package. Optimization.jl: A unified optimization package. Zenodo . APACrefURL https://doi.org/10.5281/zenodo.7738525 APACrefURL APACrefDOI doi:10.5281/zenodo.7738525 APACrefDOI

    work page doi:10.5281/zenodo.7738525 2023

  17. [17]

    , Beck, H

    FengEtAl2023 APACrefauthors Feng, D. , Beck, H. , Lawson, K. \ Shen, C. APACrefauthors \ 2023 . The suitability of differentiable, physics-informed machine learning hydrologic models for ungauged regions and climate change impact assessment The suitability of differentiable, physics-informed machine learning hydrologic models for ungauged regions and clim...

    work page doi:10.5194/hess-27-2357-2023 2023

  18. [18]

    \ Maurel, F

    goutal1997proceedings APACrefauthors Goutal, N. \ Maurel, F. APACrefauthors \ ( ). \ 1997 . Proceedings of the 2nd Workshop on Dam-Break Wave Simulation Proceedings of the 2nd workshop on dam-break wave simulation . France D\'epartement Laboratoire National d'Hydraulique, Groupe Hydraulique Fluviale Electricit\'e de France

    work page 1997

  19. [19]

    guo2024reduced APACrefauthors Guo, Q. , He, Y. , Liu, M. , Zhao, Y. , Liu, Y. \ Liu, J. APACrefauthors \ 2024 . Reduced geostatistical approach with a Fourier neural operator surrogate model for inverse modeling of hydraulic tomography Reduced geostatistical approach with a fourier neural operator surrogate model for inverse modeling of hydraulic tomograp...

    work page 2024

  20. [20]

    APACrefauthors \ 2015

    hager2015albert APACrefauthors Hager, W H. APACrefauthors \ 2015 . Albert Strickler: His life and work Albert strickler: His life and work . Journal of Hydraulic Engineering 141 7 02515002

    work page 2015

  21. [21]

    \ Syme, B

    huxley2016tuflow APACrefauthors Huxley, C. \ Syme, B. APACrefauthors \ 2016 . TUFLOW GPU-best practice advice for hydrologic and hydraulic model simulations Tuflow gpu-best practice advice for hydrologic and hydraulic model simulations . 37th Hydrology & Water Resources Symposium 37th hydrology & water resources symposium \ ( \ 195--203)

    work page 2016

  22. [22]

    APACrefauthors \ 2016

    hydronia2016riverflow2d APACrefauthors Hydronia, L L C. APACrefauthors \ 2016 . RiverFlow2D Two-dimensional River Dynamics Model, Reference Manual RiverFlow2D Two-dimensional River Dynamics Model, Reference Manual . Pembroke Pines, FL

    work page 2016

  23. [23]

    , Zheng, Y

    jiang2020improving APACrefauthors Jiang, S. , Zheng, Y. \ Solomatine, D. APACrefauthors \ 2020 . Improving AI system awareness of geoscience knowledge: Symbiotic integration of physical approaches and deep learning Improving ai system awareness of geoscience knowledge: Symbiotic integration of physical approaches and deep learning . Geophysical Research L...

    work page 2020

  24. [24]

    APACrefauthors \ 2010

    lai2010two APACrefauthors Lai, Y G. APACrefauthors \ 2010 . Two-dimensional depth-averaged flow modeling with an unstructured hybrid mesh Two-dimensional depth-averaged flow modeling with an unstructured hybrid mesh . Journal of Hydraulic Engineering 136 1 12--23

    work page 2010

  25. [25]

    , Ghorbanidehno, H

    LeeEtAl2018 APACrefauthors Lee, J. , Ghorbanidehno, H. , Farthing, M W. , Hesser, T J. , Darve, E F. \ Kitanidis, P K. APACrefauthors \ 2018 . Riverine Bathymetry Imaging With Indirect Observations Riverine bathymetry imaging with indirect observations . Water Resources Research 54 3704-3727 . APACrefDOI doi:10.1029/2017WR021649 APACrefDOI

    work page doi:10.1029/2017wr021649 2018

  26. [26]

    \ G \"u nnemann, S

    lienen2022torchode APACrefauthors Lienen, M. \ G \"u nnemann, S. APACrefauthors \ 2022 . torchode: A parallel ODE solver for pytorch torchode: A parallel ode solver for pytorch . arXiv preprint arXiv:2210.12375

    work page arXiv 2022

  27. [27]

    APACrefauthors \ 1970

    limerinos1970determination APACrefauthors Limerinos, J T. APACrefauthors \ 1970 . Determination of the Manning coefficient from measured bed roughness in natural channels Determination of the manning coefficient from measured bed roughness in natural channels . US Government Printing Office

    work page 1970

  28. [28]

    APACrefauthors \ 2025

    pyHMT2D APACrefauthors Liu, X. APACrefauthors \ 2025 . pyHMT2D User Guide . pyHMT2D User Guide . APACrefURL https://psu-efd.github.io/pyHMT2D\_API\_Web/ APACrefURL

    work page 2025

  29. [29]

    , Landry, B J

    LiuEtAl2008 APACrefauthors Liu, X. , Landry, B J. \ Garcia, M H. APACrefauthors \ 2008 . Coupled two-dimensional model for scour based on shallow water equations with unstructured mesh Coupled two-dimensional model for scour based on shallow water equations with unstructured mesh . Coastal Engineering 50 800--810

    work page 2008

  30. [30]

    , Mazdeh, A

    LiuEtAl2024NCHRP APACrefauthors Liu, X. , Mazdeh, A. , Zevenbergen, L W. \ Kramer, C M. APACrefauthors \ 2024 . NCHRP Research Report 1077: Selection and Application of Manning's Roughness Values in Two-Dimensional Hydraulic Models NCHRP research report 1077: Selection and application of Manning's roughness values in two-dimensional hydraulic models \ Nat...

    work page doi:10.17226/27730 2024

  31. [31]

    , Shen, C

    LiuEtAl2024 APACrefauthors Liu, X. , Shen, C. \ Song, Y. APACrefauthors \ 2024 . Bathymetry Inversion Using a Deep-Learning-Based Surrogate for Shallow Water Equations Solvers Bathymetry inversion using a deep-learning-based surrogate for shallow water equations solvers . Water Resources Research 60 e2023WR035890

    work page 2024

  32. [32]

    APACrefauthors \ 1891

    Manning1891 APACrefauthors Manning, R. APACrefauthors \ 1891 . On the flow of water in open channels and pipes On the flow of water in open channels and pipes . Transactions of the Institution of Civil Engineers of Ireland 20 161--207

    work page

  33. [33]

    , Roberts, K

    marcus1992evaluation APACrefauthors Marcus, W A. , Roberts, K. , Harvey, L. \ Tackman, G. APACrefauthors \ 1992 . An evaluation of methods for estimating Manning's n in small mountain streams An evaluation of methods for estimating manning's n in small mountain streams . Mountain Research and Development 227--239

    work page 1992

  34. [34]

    , Pereira, F

    meert2018surrogate APACrefauthors Meert, P. , Pereira, F. \ Willems, P. APACrefauthors \ 2018 . Surrogate modeling-based calibration of hydrodynamic river model parameters Surrogate modeling-based calibration of hydrodynamic river model parameters . Journal of Hydro-environment Research 19 56--67

    work page 2018

  35. [35]

    , Saki, S

    al2024spatiotemporal APACrefauthors Mehedi, M A. , Saki, S. , Patel, K. , Shen, C. , Cohen, S. , Smith, V. Lawson, K. APACrefauthors \ 2024 . Spatiotemporal variability of channel roughness and its substantial impacts on flood modeling errors Spatiotemporal variability of channel roughness and its substantial impacts on flood modeling errors . Earth's Fut...

    work page 2024

  36. [36]

    APACrefauthors \ 1944

    Moody1944 APACrefauthors Moody, L F. APACrefauthors \ 1944 . Friction factors for pipe flow Friction factors for pipe flow . Transactions of the American Society of Mechanical Engineers 66 671--684

    work page 1944

  37. [37]

    o mungsgesetze in rauhen R ohren Str\

    Nikuradse1933 APACrefauthors Nikuradse, J. APACrefauthors \ 1933 . Str\" o mungsgesetze in rauhen R ohren Str\" o mungsgesetze in rauhen R ohren \ Forschungsheft\ 361 . Berlin, Germany VDI-Verlag . English translation: ``Laws of flow in rough pipes,'' NACA Technical Memorandum 1292, 1950

    work page 1933

  38. [38]

    APACrefauthors \ 1950

    Nikuradse1950 APACrefauthors Nikuradse, J. APACrefauthors \ 1950 . Laws of flow in rough pipes Laws of flow in rough pipes \ Technical Memorandum\ \ 1292 . National Advisory Committee for Aeronautics . English translation of `` Str\" o mungsgesetze in rauhen Rohren ,'' VDI Forschungsheft 361, 1933

    work page 1950

  39. [39]

    APACrefauthors \ 2011

    o2011flo APACrefauthors O'Brien, J. APACrefauthors \ 2011 . FLO-2D Reference manual, version 2009 Flo-2d reference manual, version 2009 . FLO-2D Official Website. Available online: http://www. flo-2d. com (accessed on 6 June 2011)

    work page 2011

  40. [40]

    , Moteki, D

    ohara2024physics APACrefauthors Ohara, Y. , Moteki, D. , Muramatsu, S. , Hayasaka, K. \ Yasuda, H. APACrefauthors \ 2024 . Physics-informed neural networks for inversion of river flow and geometry with shallow water model Physics-informed neural networks for inversion of river flow and geometry with shallow water model . Physics of Fluids 36 10

    work page 2024

  41. [41]

    \ Ruthotto, L

    onken2020discretize APACrefauthors Onken, D. \ Ruthotto, L. APACrefauthors \ 2020 . Discretize-optimize vs. optimize-discretize for time-series regression and continuous normalizing flows Discretize-optimize vs. optimize-discretize for time-series regression and continuous normalizing flows . arXiv preprint arXiv:2005.13420

    work page arXiv 2020

  42. [42]

    rackauckas2020universal APACrefauthors Rackauckas, C. , Ma, Y. , Martensen, J. , Warner, C. , Zubov, K. , Supekar, R. Ramadhan, A. APACrefauthors \ 2020 . Universal Differential Equations for Scientific Machine Learning Universal differential equations for scientific machine learning . arXiv preprint arXiv:2001.04385

    work page internal anchor Pith review Pith/arXiv arXiv 2020

  43. [43]

    \ Nie, Q

    rackauckas2017differentialequations APACrefauthors Rackauckas, C. \ Nie, Q. APACrefauthors \ 2017 . Differentialequations.jl--a performant and feature-rich ecosystem for solving differential equations in julia Differentialequations.jl--a performant and feature-rich ecosystem for solving differential equations in julia . Journal of Open Research Software 5 1 15

    work page 2017

  44. [44]

    , Appling, A

    rahmani2023identifying APACrefauthors Rahmani, F. , Appling, A. , Feng, D. , Lawson, K. \ Shen, C. APACrefauthors \ 2023 . Identifying structural priors in a hybrid differentiable model for stream water temperature modeling Identifying structural priors in a hybrid differentiable model for stream water temperature modeling . Water Resources Research 59 12...

    work page 2023

  45. [45]

    , Perdikaris, P

    raissi2019physics APACrefauthors Raissi, M. , Perdikaris, P. \ Karniadakis, G E. APACrefauthors \ 2019 . Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations Physics-informed neural networks: A deep learning framework for solving forward and inverse problems...

    work page 2019

  46. [46]

    , Borthwick, A G

    RogersEtAl2003 APACrefauthors Rogers, B. , Borthwick, A G. \ Taylor, P H. APACrefauthors \ 2003 . Mathematical balancing of flux gradient and source terms prior to using Roe's approximate Riemann solver Mathematical balancing of flux gradient and source terms prior to using roe's approximate riemann solver . Journal of Computational Physics 192 2 422-451 ...

    work page doi:10.1016/j.jcp.2003.07.020 2003

  47. [47]

    , Fujihara, M

    RogersEtAl2001 APACrefauthors Rogers, B. , Fujihara, M. \ Borthwick, A G L. APACrefauthors \ 2001 . Adaptive Q-tree Godunov-type scheme for shallow water equations Adaptive q-tree godunov-type scheme for shallow water equations . International Journal for Numerical Methods in Fluids 35 3 247-280 . APACrefURL https://onlinelibrary.wiley.com/doi/abs/10.1002...

    work page doi:10.1002/1097-0363 2001

  48. [48]

    APACrefauthors \ 1965

    Rouse1965 APACrefauthors Rouse, H. APACrefauthors \ 1965 . Critical analysis of open-channel resistance Critical analysis of open-channel resistance . Journal of the Hydraulics Division 91 HY4 1--25

    work page 1965

  49. [49]

    , Bolibar, J

    sapienza2024differentiable APACrefauthors Sapienza, F. , Bolibar, J. , Schäfer, F. , Groenke, B. , Pal, A. , Boussange, V. Rackauckas, C. APACrefauthors \ 2024 . Differentiable Programming for Differential Equations: A Review. Differentiable programming for differential equations: A review

    work page 2024

  50. [50]

    APACrefauthors \ 1991

    schiesser1991method APACrefauthors Schiesser, W E. APACrefauthors \ 1991 . The Numerical Method of Lines: Integration of Partial Differential Equations The numerical method of lines: Integration of partial differential equations . Academic Press

    work page 1991

  51. [51]

    , Appling, A P

    Shen2023 APACrefauthors Shen, C. , Appling, A P. , Gentine, P. , Bandai, T. , Gupta, H. , Tartakovsky, A. Lawson, K. APACrefauthors \ 2023 . Differentiable modelling to unify machine learning and physical models for geosciences Differentiable modelling to unify machine learning and physical models for geosciences . Nature Reviews Earth & Environment 4 552...

    work page doi:10.1038/s43017-023-00450-9 2023

  52. [52]

    \ Phanikumar, M S

    Shen2010 APACrefauthors Shen, C. \ Phanikumar, M S. APACrefauthors \ 2010 . A process-based, distributed hydrologic model based on a large-scale method for surface–subsurface coupling A process-based, distributed hydrologic model based on a large-scale method for surface–subsurface coupling . Advances in Water Resources 33 1524-1541 . APACrefDOI doi:10.10...

    work page doi:10.1016/j.advwatres.2010.09.002 2010

  53. [53]

    , Le Maitre, O

    siripatana2020bayesian APACrefauthors Siripatana, A. , Le Maitre, O. , Knio, O. , Dawson, C. \ Hoteit, I. APACrefauthors \ 2020 . Bayesian inference of spatially varying Manning’sn coefficients in an idealized coastal ocean model using a generalized Karhunen-Lo \`e ve expansion and polynomial chaos Bayesian inference of spatially varying manning’sn coeffi...

    work page 2020

  54. [54]

    , Mayo, T

    siripatana2018ensemble APACrefauthors Siripatana, A. , Mayo, T. , Knio, O. , Dawson, C. , Le Ma \^ tre, O. \ Hoteit, I. APACrefauthors \ 2018 . Ensemble Kalman filter inference of spatially-varying Manning’sn coefficients in the coastal ocean Ensemble kalman filter inference of spatially-varying manning’sn coefficients in the coastal ocean . Journal of Hy...

    work page 2018

  55. [55]

    , Bindas, T

    song2024high APACrefauthors Song, Y. , Bindas, T. , Shen, C. , Ji, H. , Knoben, W J M. , Lonzarich, L. others APACrefauthors \ 2024 . High-resolution national-scale water modeling is enhanced by multiscale differentiable physics-informed machine learning High-resolution national-scale water modeling is enhanced by multiscale differentiable physics-informe...

    work page 2024

  56. [56]

    , Knoben, W J

    song2024ancient APACrefauthors Song, Y. , Knoben, W J. , Clark, M P. , Feng, D. , Lawson, K. , Sawadekar, K. \ Shen, C. APACrefauthors \ 2024 . When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling When ancient numerical demons meet physics-informed machine learning: adjoint-base...

    work page 2024

  57. [57]

    , Sawadekar, K

    song2024improving APACrefauthors Song, Y. , Sawadekar, K. , Frame, J M. , Pan, M. , Clark, M. , Knoben, W J. Shen, C. APACrefauthors \ 2024 . Improving Physics-informed, Differentiable Hydrologic Models for Capturing Unseen Extreme Events Improving physics-informed, differentiable hydrologic models for capturing unseen extreme events . Authorea Preprints

    work page 2024

  58. [58]

    , Shen, C

    song2023surrogate APACrefauthors Song, Y. , Shen, C. \ Liu, X. APACrefauthors \ 2023 . A Surrogate Model for Shallow Water Equations Solvers with Deep Learning A surrogate model for shallow water equations solvers with deep learning . Journal of Hydraulic Engineering 149 04023045

    work page 2023

  59. [59]

    APACrefauthors \ 2022

    TRITON2022 APACrefauthors TRITON . APACrefauthors \ 2022 . TRITON . TRITON . APACrefURL https://triton.ornl.gov/ APACrefURL [Accessed 01-December-2023]

    work page 2022

  60. [60]

    Runge–Kutta pairs of order 5(4) satisfying only the first column simplifying assumption

    Tsitouras2011 APACrefauthors Tsitouras, C. APACrefauthors \ 2011 . Runge–Kutta pairs of order 5(4) satisfying only the first column simplifying assumption Runge–kutta pairs of order 5(4) satisfying only the first column simplifying assumption . Computers & Mathematics with Applications 62 2 770-775 . APACrefURL https://www.sciencedirect.com/science/articl...

    work page doi:10.1016/j.camwa.2011.06.002 2011

  61. [61]

    , Younis, J

    van2010lisflood APACrefauthors Van Der Knijff, J. , Younis, J. \ De Roo, A. APACrefauthors \ 2010 . LISFLOOD: a GIS-based distributed model for river basin scale water balance and flood simulation Lisflood: a gis-based distributed model for river basin scale water balance and flood simulation . International Journal of Geographical Information Science 24 ...

    work page 2010

  62. [62]

    , Jiang, S

    wang2024distributed APACrefauthors Wang, C. , Jiang, S. , Zheng, Y. , Han, F. , Kumar, R. , Rakovec, O. \ Li, S. APACrefauthors \ 2024 . Distributed hydrological modeling with physics-encoded deep learning: A general framework and its application in the Amazon Distributed hydrological modeling with physics-encoded deep learning: A general framework and it...

    work page 2024

  63. [63]

    \ Bach, H

    warren1992mike APACrefauthors Warren, I. \ Bach, H. APACrefauthors \ 1992 . MIKE 21: a modelling system for estuaries, coastal waters and seas Mike 21: a modelling system for estuaries, coastal waters and seas . Environmental Software 7 4 229--240

    work page 1992

  64. [64]

    , Zhang, J

    wei2024effects APACrefauthors Wei, Z. , Zhang, J. , Wang, D. , Gao, Y. \ Cheng, J. APACrefauthors \ 2024 . The effects of non-local observations on the adjoint estimation of local model parameters: An example of Manning’sn coefficient in a tidal model over the Bohai, Yellow, and East China Seas The effects of non-local observations on the adjoint estimati...

    work page 2024

  65. [65]

    , Lehman, W

    wing2022inequitable APACrefauthors Wing, O E. , Lehman, W. , Bates, P D. , Sampson, C C. , Quinn, N. , Smith, A M. Kousky, C. APACrefauthors \ 2022 . Inequitable patterns of US flood risk in the Anthropocene Inequitable patterns of us flood risk in the anthropocene . Nature Climate Change 12 2 156--162

    work page 2022

  66. [66]

    APACrefauthors \ 1992

    yen1992dimensionally APACrefauthors Yen, B C. APACrefauthors \ 1992 . Dimensionally homogeneous Manning's formula Dimensionally homogeneous manning's formula . Journal of hydraulic engineering 118 9 1326--1332

    work page 1992

  67. [67]

    APACrefauthors \ 2002

    Yen2002 APACrefauthors Yen, B C. APACrefauthors \ 2002 . Open channel flow resistance Open channel flow resistance . Journal of Hydraulic Engineering 128 1 20--39

    work page 2002

  68. [68]

    , Lei, H

    zhong2024development APACrefauthors Zhong, L. , Lei, H. \ Yang, J. APACrefauthors \ 2024 . Development of a distributed physics-informed deep learning hydrological model for data-scarce regions Development of a distributed physics-informed deep learning hydrological model for data-scarce regions . Water Resources Research 60 6 e2023WR036333

    work page 2024