The Geometry of Flow: Advancing Predictions of River Geometry With Multi‐Model Machine Learning
Open accessThe DOI matches its Crossref record; 71 of the 71 references it cites were found in a registry or an index; 1 of the 34 sources its claims cite were checked for support.
Abstract
Hydraulic geometry parameters describing river hydrogeomorphic relationships are critical for determining a channel's capacity to convey water and sediment which is important for flood forecasting. Although well‐established, power‐law hydraulic geometry curves have been widely used to understand riverine systems and mapping flooding inundation worldwide for the past 70 years, we have become increasingly aware of their limitations. In the present study, we have moved beyond these traditional power‐law relationships, testing the ability of machine‐learning models to provide improved predictions of river width and depth. For this work, we have used an unprecedentedly large river measurement data set (HYDRoSWOT) as well as a suite of watershed predictor data to develop novel data‐driven approaches to better estimate river geometries over the contiguous United States (CONUS). Our Random Forest, XGBoost, and neural network models out‐performed the traditional, regionalized power law‐based hydraulic geometry equations for both width and depth, providing R‐squared values of as high as 0.75 for width and as high as 0.67 for depth, compared with R‐squared values of 0.45 for width and 0.18 for depth from the regional hydraulic geometry equations. Our results also show diverse performance outcomes across stream orders and geographical regions for the different machine‐learning models, demonstrating the value of using multi‐model approaches to maximize the predictability of river geometry. The developed models have been used to create the newly publicly available STREAM‐geo data set, which provides river width, depth, width/depth ratio, and river and stream surface area (%RSSA) for nearly 2.7 million NHDPlus stream reaches across the contiguous US.
Overview
CoChat Summary
TL;DR
This study tests multi-model machine learning to predict river width and depth across the contiguous US using HYDRoSWOT observations and 17 watershed predictors, finding that ML models outperform traditional regional power-law hydraulic geometry. It releases STREAM-geo predictions for nearly 2.7 million NHDPlus stream reaches.
Key findings
- Random Forest, XGBoost, and neural network (MLP) models outperformed regional power-law equations for median river width and depth. Width: RF R² = 0.75, XGBoost R² = 0.73, MLP R² = 0.73; depth: RF R² = 0.67, XGBoost R² = 0.67, MLP R² = 0.66. Regional power-law equations gave R² = 0.45 for width and R² = 0.18 for depth; multiple linear regression was worse for width (R² = 0.33).
- Median river discharge was the most important predictor in the Random Forest model (RFI width = 0.67, RFI depth = 0.78), with catchment area second-most important for width. Simple regressions showed discharge explained width R² = 0.61 and depth R² = 0.57; catchment area explained width R² = 0.36 and depth R² = 0.25; upstream dams explained width R² = 0.21 and depth R² = 0.16, though dam RFI was only 0.01.
- Performance varied by stream order and ecoregion. Power-law models overall underpredicted width by about 24% and depth by about 22%; ML models had overall under/overprediction of 7% or less for width and 13% or less for depth. For width, power-law performed best at 4th- and 5th-order streams (PBIAS = 8.9% and 4.6%) but underpredicted higher-order widths by 50% or more. For depth, power-law overpredicted 4th-order streams (median PBIAS = 8.5%) but underpredicted 1st-order (22.5%) and 9th-order (54.23%) streams. For 1st-order width, median PBIAS was RF = 22.6%, XGBoost = 30.0%, MLP = 28.9%. All ML width models performed well in Southeastern US plains (ER 8.3, NSE > 0.84), while RF vs. MLP differences reversed by region and by width vs. depth.
- STREAM-geo provides predictions for 2,610,904 of 2,691,339 NHDPlus reaches, with remaining reaches marked NaN. It includes river width, depth, width/depth ratio, and %RSSA. Across CONUS, streams and rivers cover nearly 75,000 km², or 0.93% of the continental US. Median width/depth ratio ranged from 13.0 for 1st-order streams to 48.4 for 10th-order streams, plateauing above 8th order. More than 96.9% of US stream reaches are narrower than 50 m.
Methods
- Data: HYDRoSWOT USGS ADCP observations (220,000+ records, >5,000 stations, 1940s–2014). Cleaning removed per-site width/depth values above the 95th or below the 5th percentile, NaNs, and sites with fewer than 6 data points; median width and depth were computed per site. Final training data included 5,269 reaches for width and 2,678 reaches for depth. Seventeen stream and catchment predictors were used, including climate, streamflow, NHDPlus reach metadata, bed D50, land use, vegetation, soils, population density, and dams; HYDRoSWOT locations were matched to the closest NHDPlus reach within a 500-m buffer.
- Models: regional power-law equations for 20 Hydrologic Landscape Regions, multiple linear regression, Random Forest, XGBoost, and multilayer perceptron (MLP). Data were split 75:25 training/testing. Evaluation metrics were RMSE, PBIAS, R², and NSE. Hyperparameters were optimized with GridSearchCV and 5-fold cross-validation; the MLP used 100 hyperparameter sets, Adam optimization, MSE loss, and early stopping. Random Forest predictor importance was quantified by permutation-based MSE reduction.
Limitations
- The paper states that HYDRoSWOT sampling is uneven, with more intensive sampling in the eastern US and more width than depth data; training data disproportionately represent 4th- and 5th-order streams while 1st-order streams are under-sampled. Depth predictions were less accurate than width predictions, multiple linear regression performed poorly for width, and reasons for ecoregion differences were not fully understood. Bedrock data are unavailable at national scale, so slope and elevation act as proxies. SWOT observes only rivers wider than 100 m, potentially as small as 50 m, and has limited temporal sampling, while >96.9% of US reaches are narrower than 50 m.
- Evident from the text: the models predict median width and depth under median-flow conditions, not full distributions or event-based dynamics. Performance varies substantially by stream order and region, so a single model may be unreliable, especially for low-order streams.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures

















Verification
Claims the paper makes
0 Valid · 0 Not valid · 19 Not checked · 64 Own results and interpretations
None of the sources cited for this claim have been checked yet.
In the paper
“Stream order also has strong explanatory power, which is to be expected due to the strong correlation between stream order and catchment area (Downing, 2010;Lin et al., 2020).”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 3.1. Key Predictors of River Width and Depth
What the sources say
- Could not verifycites Downing, 2010 matched to “Emerging global role of small lakes and ponds: little things mean a lot” (2010) sourceExperimental — accuracy not yet measured
Our copy of Downing (2010, Limnetica) is entirely about the global abundance, size distribution and biogeochemical role of small lakes and ponds (Pareto size distributions, 'Why might small lakes and ponds be very important?', 'Lakes, ponds, rivers, and streams are of global importance'); the only river/stream content is the generic 'neutral pipe' vs 'active pipe' carbon discussion. No passage mentions stream order, Strahler ordering, or a correlation between stream order and catchment/drainage area, nor river width or depth as predictands (searched: 'stream order Strahler catchment watershed area rivers', 'river width depth drainage basin area streams', 'number and area of streams and rivers'). Since our copy neither states nor contradicts the attributed correlation, the claim cannot be checked against this source. (LLM, conf=0.68)
- Not checked yetcites Lin et al., 2020 matched to “Global Estimates of Reach‐Level Bankfull River Width Leveraging Big Data Geospatial Analysis” (2020) source
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is important to note here that the high level of performance with the current modeling approaches was achieved using only publicly available land surface information, without the need for detailed subsurface altimetry and satellite information.”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 4. Conclusions
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Another important predictor of both width and depth is the number of upstream dams, with more dams being associated with wider and deeper streams (width, R 2 = 0.21; depth, R 2 = 0.16).”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 3.1. Key Predictors of River Width and Depth
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“All machine-learning models provided exceptional evaluation metrics that outperformed state-of-the-art regional (physiographic) hydraulic geometry equations, developed across hydrological landscape regions (HLRs), which have been widely used in large-scale hydrology models for flood prediction and inundation mapping, including the National Water Model.”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 4. Conclusions
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Finally, the results of the present work have resulted in the development of STREAM-geo, a comprehensive river geometry database for the continental United States.”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 4. Conclusions
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
35 more claims
Create a free account to see all 83 claims and to verify the 27 references still open.
- It may state two separate claims as one. 33
- Probably not a substantial claim. 14
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 3.5. Stream Reach and Metrics-Geometry (STREAM-geo)
“As seen in Figure 10b, the width-to-depth ratio varies greatly across stream orders, ranging from a median value of 13.0 for 1st-order streams to 48.4 for 10th-order streams.”
§ The Geometry of Flow: Advancing Predictions of River Geometry With Multi-Model Machine Learning > 4. Conclusions
“(b) River width/depth distributions across the CONUS.(c) River width/depth ratios across stream orders.”
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016
Multilayer feedforward networks are universal approximators
Hornik, K., Stinchcombe, M., & White, H. (1989) · Neural Networks · doi:10.1016/0893-6080(89)90020-8 · Cited by 21,728 · cited 1×
“In the current study, we employed a multilayer perceptron (MLP), one of the most popular neural network architectures (Hornik et al., 1989).”
Random forests for classification in ecology
Cutler, D. R., Edwards, T. C., Beard, K. H., Cutler, A., Hess, K. T., Gibson, J., & Lawler, J. J. (2007) · Ecology · doi:10.1890/07-0539.1 · Cited by 4,950
State-of-the-art in artificial neural network applications: A survey
Abiodun, O. I., Jantan, A., Omolara, A. E., Dada, K. V., Mohamed, N. A., & Arshad, H. (2018) · Heliyon · doi:10.1016/j.heliyon.2018.e00938 · Cited by 3,054
The hydraulic geometry of stream channels and some physiographic implications
Leopold, L. B., & Maddock, T. (1953) · Professional Paper · doi:10.3133/pp252 · Cited by 2,966 · cited 4×
“In addition, regional (physiographic) hydraulic curves (Blackburn‐Lynch et al., 2017), similar to hydraulic geometry curves (Leopold & Maddock, 1953), which demonstrate a power‐law relationship between the catchment drainage area and the bankfull characteristics, are used to determine the bankfull width and cross‐sect…”
InfluentialFlood risk and climate change: global and regional perspectives
Kundzewicz, Z. W., Kanae, S., Seneviratne, S. I., Handmer, J., Nicholls, N., Peduzzi, P., Mechler, R., Bouwer, L. M., Arnell, N., Mach, K., et al. (2014) · Hydrological Sciences Journal · doi:10.1080/02626667.2013.857411 · Cited by 1,651
A review of the potential impacts of climate change on surface water quality
WHITEHEAD, P. G., WILBY, R. L., BATTARBEE, R. W., KERNAN, M., & WADE, A. J. (2009) · Hydrological Sciences Journal · doi:10.1623/hysj.54.1.101 · Cited by 1,305
Version 3 of the Global Aridity Index and Potential Evapotranspiration Database
Zomer, R. J., Xu, J., & Trabucco, A. (2022) · Scientific Data · doi:10.1038/s41597-022-01493-1 · Cited by 1,263
Shen, C. (2018) · Water Resources Research · doi:10.1029/2018wr022643 · Cited by 1,133
Self-citation
Research works printed but not linked: 3
- Chollet, F. (2021). Deep learning with Python (2nd ed.), Manning Publications.
- Cutler, D. R., Edwards, T. C., Jr., Beard, K. H., Cutler, A., Hess, K. T., Gibson, J., & Lawler, J. J. (2007). Random forests for classification in ecology. Ecology, 88(11), 2783-2792. https://doi.org/10.1890/07-0539. · doi:10.1890/07-0539
- 10.1061/(asce)0733‐9429(1983)109:8(1073 · doi:10.1061/(asce)0733-9429(1983)109:8(1073
Cited by
Explainable artificial intelligence for sustainable urban water systems engineering
Infant, S. S., Vickram, S., Saravanan, A., Muthu, C. M. M., & Yuarajan, D. (2025) · Results in Engineering · doi:10.1016/j.rineng.2025.104349 · Cited by 107
Distinct hydrologic response patterns and trends worldwide revealed by physics-embedded learning
Ji, H., Song, Y., Bindas, T., Shen, C., Yang, Y., Pan, M., Liu, J., Rahmani, F., Abbas, A., Beck, H. E., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-64367-1 · Cited by 39
Self-citationTikuye, B. G., Ray, R. L., Abeysingha, N. S., & Gurau, S. (2025) · Progress in Disaster Science · doi:10.1016/j.pdisas.2025.100462 · Cited by 22
Zarrabi, R., McDermott, R., Erfani, S. M. H., & Cohen, S. (2025) · Water Resources Research · doi:10.1029/2024wr037997 · Cited by 13
Self-citationGao, W., Li, F., Cai, Y., & Hou, X. (2025) · Journal of Hydrology Regional Studies · doi:10.1016/j.ejrh.2025.102799 · Cited by 8
Physics-informed neural network for open channel flow velocity prediction
Prieto, C., Huang, J., Ansari, S., & Schell, K. R. (2025) · Physics of Fluids · doi:10.1063/5.0273148 · Cited by 6
White, D. C., Morrison, R. R., & Nelson, P. A. (2025) · Journal of Geophysical Research Earth Surface · doi:10.1029/2024jf008038 · Cited by 5
Qiao, H., Jooshaki, M., Rolle, M., Lähivaara, T., Vauhkonen, M. J., Kauppila, T., & Muniruzzaman, M. (2025) · Water Resources Research · doi:10.1029/2024wr039487 · Cited by 5
High‐Resolution Channel Geometry Reveals Contrasting Styles of Gravel River Adjustment
Kostynick, R. P., Phillips, C. B., & Masteller, C. C. (2026) · Geophysical Research Letters · doi:10.1029/2025gl118412 · Cited by 4
Redefining Bankfull Stage Using Machine Learning Prediction in a Large Lowland River
Amanambu, A. C., Mossa, J., & Chen, Y. (2025) · Earth Surface Processes and Landforms · doi:10.1002/esp.70211 · Cited by 1
Zhang, L., Borse, D., Singh, A., Pizzuto, J. E., Fu, X., & Parker, G. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2513733123 · Cited by 1
Malekpour, M. M., Ahmadi, M. M., & Qaderi, K. (2025) · International Journal of Environmental Research · doi:10.1007/s41742-025-00922-4 · Cited by 1
Malekpour, M. M., Ahmadi, M. M., Qaderi, K., & Rajabizadeh, Y. (2024) · Industrial Innovations · doi:10.61882/jii.2.3.285 · Cited by 1
Estimation of spatiotemporally varying parameters for grid-based distributed hydrologic models
Zhang, X., Liu, P., Xie, K., Liu, W., Deng, L., Xu, H., Cheng, Q., & Zhou, L. (2025) · Journal of Hydrology Regional Studies · doi:10.1016/j.ejrh.2025.102536
Castejón-Villalobos, J., Lee, A., Patterson, N. K., Lane, B., & Phillips, C. B. (2026) · Water Resources Research · doi:10.1029/2025wr042220
Deep Learning Recognition and Prediction Algorithms for Channel Evolution in the Lower Yellow River
Chunlong, S. (2025) · International Conference on Computers, Information Processing and Advanced Education (CIPAE) · doi:10.1109/cipae66821.2025.00192
Evaluation of Optimized Zayandehrud River Channel Dimensions to Mitigate Flood Damage
Malekpour, M. M., Ahmadi, M. M., & Qaderi, K. (2025) · CLEAN - Soil Air Water · doi:10.1002/clen.70074
Yang, Y., Chen, C., Yang, G., Sheng, H., Shi, Y., Lyu, J., Liu, T., Liu, S., He, C., Ou, Z., et al. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5527811
Geometric representation of rivers and lakes in hydrologic models: a reviewPreprint
Liao, C., Zhou, T., Fang, Y., Craig, J. R., Duan, Z., Sun, N., Yang, Z., Tan, Z., Hetland, R. D., Li, H. Y., et al. (2026) · doi:10.22541/essoar.176805006.68588074/v1
Malekpour, M. M., Ahmadi, M. M., Qaderi, K., Gugliotta, M., & Rajabizadeh, Y. (2026) · Journal of Hydrologic Engineering · doi:10.1061/jhyeff.heeng-6567
Liu, Q., Zheng, J., Zheng, C., Zhao, F., Yu, L., Ling, F., Da, Z., & Lian, J. (2026) · Water Resources Management · doi:10.1007/s11269-026-04642-y
A hybrid machine learning and physics-based approach to model high-resolution inundation
Peeples, A., Leonarduzzi, E., Condon, L., & Maxwell, R. M. (2026) · Journal of Hydrology · doi:10.1016/j.jhydrol.2026.136246
Process from Noise: Erosion Threshold Variability and the Morphology of Gravel-Bed Rivers
Kostynick, R. P. (2026) · Open Scholarship Institutional Repository (Washington University in St. Louis) · doi:10.7936/ewy7-ey88
Metadata
- Authors
- Chang, S. Y., Ghahremani, Z., Manuel, L., Erfani, S. M. H., Shen, C., Cohen, S., Van Meter, K. J., Pierce, J. L., Meselhe, E. A., & Goharian, E.
- Affiliations
- Pennsylvania State University (US)
- Boise State University (US)
- Tulane University (US)
- University of South Carolina (US)
- University of Alabama (US)
- Department of Geography Pennsylvania State University University Park PA USA
- Corresponding author
- Shen, Chaopeng
- Cohen, Sagy
- ORCID
- Venue
- Water Resources Research
- Publisher
- American Geophysical Union (AGU)
- Published
- 2024-10
- Volume
- 60
- Issue
- 10
- ISSN
- 0043-1397 (ISSN-L)
- 1944-7973
- OpenAlex
- W4403686990
- Semantic Scholar
- a6b958022ab2c9a3ba8ae388c86eaa9a19e4c29e
- Cited by
- 23
- References
- 69
- Citations by year
- 2026: 8 · 2025: 14 · 2024: 1
- FWCI
- 4.98
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Water Resources Research (publishedVersion, cc-by-nc-nd)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Consortium of Universities for the Advancement of Hydrologic Science
- National Oceanic and Atmospheric Administration (A22-0307-S003)
- National Water Center, United Arab Emirates University
- Topics
- Physical Sciences › Earth and Planetary Sciences › Earth-Surface Processes › Hydrology and Sediment Transport Processes
- Physical Sciences › Environmental Science › Global and Planetary Change › Flood Risk Assessment and Management
- Physical Sciences › Environmental Science › Water Science and Technology › Hydrology and Watershed Management Studies
- Keywords
- river geometry, hydraulic geometry, river width, river depth, power-law relationship, multi-model ensemble, random forest, XGBoost, artificial neural networks, flood inundation, contiguous United States
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-11-19, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-11-19 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.