Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 85 of the 88 references it cites were found in a registry or an index; none of the 34 sources its claims cite is checked for support yet.

A Comprehensive Assessment of Floodwater Depth Estimation Models in Semiarid Regions

Open access

Abstract

Simple models continue to be important for continental‐scale floodwater depth mapping due to the prohibitively expensive cost of calibrating and applying hydrodynamic models. This paper investigates the accuracy of three simple models for floodwater depth estimation from remote sensing derived water extent and/or Digital Elevation Models (DEMs) in semiarid regions. The three models are Height Above Nearest Drainage (HAND; Nobre et al., 2011, https://doi.org/10.1016/j.jhydrol.2011.03.051 ), Teng Vaze Dutta (TVD; Teng et al., 2013, http://hdl.handle.net/102.100.100/97033?index=1 ), and Floodwater Depth Estimation Tool (FwDET; Cohen, Brakenridge, et al., 2018, https://doi.org/10.1111/1752-1688.12609 ). The model accuracy and nature of errors are established using industry's best practice hydrodynamic models as benchmarks in three regions in eastern Australia. The overall results show that FwDET tends to underestimate (by 0.32 m at 50th percentile) while HAND and TVD overestimate floodwater depth for almost all floods (by 0.97 and 0.98 m, respectively). We quantify how switching DEM from 5 m LiDAR to national or global data sets DEM‐H (Gallant et al., 2011, https://ecat.ga.gov.au/geonetwork/srv/eng/catalog.search#/metadata/72759 ), MERIT (Yamazaki et al., 2019, https://doi.org/10.1029/2019WR024873 ), or FABDEM (Hawker et al., 2022, https://doi.org/10.1088/1748-9326/ac4d4f ) can affect different models differently; and we evaluate model performance against reach geomorphology and magnitude of flood events. The findings emphasize the importance of choosing a model that is fit for the intended application. By describing the applicability, advantages, and limitations of these models, this paper assists practitioners to choose the most appropriate model based on characteristics of their study area, type of problems they try to solve, and data availability.

Overview

CoChat Summary

TL;DR

This paper evaluates three simple remote-sensing/DEM-based floodwater depth models—HAND, TVD, and FwDET—against hydrodynamic benchmark models across semiarid river-floodplain systems in Australia’s Murray-Darling Basin. FwDET had the lowest overall error, but model choice depends on application, available inputs, and flood depth.

Key findings

  • Across 26 flood scenes, FwDET had the lowest median MAE (0.57 m) and RMSE (0.78 m), followed by TVD (MAE 1.04 m) and HAND (MAE 1.38 m). FwDET generally underestimated depth (23 of 26 scenes; median underprediction 0.32 m), while HAND overestimated in 19 of 26 scenes (median 0.97 m) and TVD overestimated in 20 of 26 scenes (median 0.98 m).
  • FwDET was most accurate for shallow water: median MAE 0.49 m for depths below 2 m, but its error increased to 1.41 m for depths above 4 m. For depths above 4 m, TVD had a median MAE of 1.35 m and HAND 1.53 m; all three models performed similarly for depths between 2 and 4 m.
  • Using national/global DEMs instead of 5 m LiDAR degraded accuracy least for FwDET (median MAE 0.57 to 0.65–0.72 m), followed by TVD, while HAND degraded most (38%–114% increase). FABDEM performed best with FwDET and TVD; DEM-H performed best with HAND.
  • FwDET errors were reasonably consistent across scenes and fitted a Cauchy distribution (x0 = 0.31, γ = 0.30), whereas HAND and TVD error distributions were too flood- and location-specific to generalize. Model accuracy improved only weakly with larger return intervals (R² 0.01–0.16), and MAE increased with mean bank-full width for all models (R² 0.19–0.46).

Methods

  • The study assessed HAND, TVD, and FwDET in three semiarid Murray-Darling Basin river-floodplain systems: the Namoi River, the River Murray in South Australia, and the Balonne/Lower Balonne system. These covered 11 reaches, 4,350 km², 7 flood events with peak discharges of 25–300 GL/d, and 26 reach-scene combinations (“26 flood scenes”).
  • Benchmark data were outputs from three calibrated hydrodynamic models, plus a 5 m LiDAR DEM and three national/global DEMs: DEM-H (~30 m), FABDEM (~30 m), and MERIT (~90 m). Observed flood extents came from Landsat 5/7/8 imagery using a multi-index method and aerial photography; water levels for HAND were extracted from hydrodynamic models upstream and downstream of each reach.
  • The simple models were run with their required inputs: HAND used DEM plus water level; TVD used DEM plus flood extent with a slope-adjusted DEM; FwDET used DEM plus flood extent and perimeter-based water-surface interpolation. Accuracy was evaluated with RMSE, MAE, and percentiles, with bootstrapping to 10,000 samples, flood-frequency analysis using Log-Pearson III, and correlation with reach geomorphology.

Limitations

  • Stated: The authors note uncertainty in how they applied the models and therefore do not report a ranking based on errors alone. TVD’s slope-defining points were selected subjectively; HAND is sensitive to stream-network definition and used a trial-and-error accumulation threshold; FwDET relies on interpolation and becomes less accurate away from inundation boundaries. They also note that comparison was limited to dates with high-quality flood-extent imagery, which may favor TVD and FwDET, and that HAND used hydrodynamic-modeled water levels rather than hydrological-model-derived levels, which would add uncertainty in real applications.
  • Stated: Return-interval analysis was weak (R² 0.01–0.16), bed-roughness data were insufficient to quantify differences between reaches, and it was difficult to select a model based on reach/floodplain characteristics.
  • Evident from the text: The assessment is limited to three semiarid Murray-Darling Basin study areas and 26 flood scenes, with benchmark hydrodynamic models that themselves carry error; DEM issues include a false channel bottom in the Namoi LiDAR DEM, poorly captured bathymetry around a weir pool, and Landsat artifacts/cloud and ETM+ scan-line issues in South Australia. The simple models also cannot predict velocity.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1029/2022wr032031

Cites

85 of 85 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 The study sites for the model comparison included four reaches of the (a) Namoi River, three reaches of the (b) River Murray, and four reaches of the (c) Balonne River. Background imagery: Bing Maps Aerial © 2022 Microsoft Corporation Earthstar Geographics SIO.
Figure 1 The study sites for the model comparison included four reaches of the (a) Namoi River, three reaches of the (b) River Murray, and four reaches of the (c) Balonne River. Background imagery: Bing Maps Aerial © 2022 Microsoft Corporation Earthstar Geographics SIO.CC-BY · From the PDF
Image 2
CC-BY · From the PDF
Figure 2 Conceptual models and processing steps for deriving depth from Height Above Nearest Drainage (HAND), Teng Vaze Dutta (TVD), and Floodwater Depth Estimation Tool (FwDET). HAND uses observed levels and a Digital Elevation Model (DEM) to calculate flood depth at address points. TVD uses an inundation extent raster and a slope-adjusted DEM to calculate the maximum surface water level across inundated areas. FwDET uses an inundation extent raster and a DEM to calculate depth by extrapolating between surface water levels identified along the perimeter of an inundation.
Figure 2 Conceptual models and processing steps for deriving depth from Height Above Nearest Drainage (HAND), Teng Vaze Dutta (TVD), and Floodwater Depth Estimation Tool (FwDET). HAND uses observed levels and a Digital Elevation Model (DEM) to calculate flood depth at address points. TVD uses an inundation extent raster and a slope-adjusted DEM to calculate the maximum surface water level across inundated areas. FwDET uses an inundation extent raster and a DEM to calculate depth by extrapolating between surface water levels identified along the perimeter of an inundation.CC-BY · From the PDF
Image 4
CC-BY · From the PDF
Figure 3 Difference between Height Above Nearest Drainage (blue), Teng Vaze Dutta (orange), and Floodwater Depth Estimation Tool (green) predictions and hydrodynamic model outputs (benchmark) for the 26 flood scenes. For each combination, the median is shown with a horizontal line, the boxes extend from 25th to 75th percentile and the whiskers from 2.5th to 97.5th percentile.
Figure 3 Difference between Height Above Nearest Drainage (blue), Teng Vaze Dutta (orange), and Floodwater Depth Estimation Tool (green) predictions and hydrodynamic model outputs (benchmark) for the 26 flood scenes. For each combination, the median is shown with a horizontal line, the boxes extend from 25th to 75th percentile and the whiskers from 2.5th to 97.5th percentile.CC-BY · From the PDF
Figure 4 Comparison of the spatial variation in errors for Height Above Nearest Drainage, Floodwater Depth Estimation Tool, and Teng Vaze Dutta along the Namoi River on 24 July 1998. Maps of the other floods are provided in Figures S5-S30 of Supporting Information S1.
Figure 4 Comparison of the spatial variation in errors for Height Above Nearest Drainage, Floodwater Depth Estimation Tool, and Teng Vaze Dutta along the Namoi River on 24 July 1998. Maps of the other floods are provided in Figures S5-S30 of Supporting Information S1.CC-BY · From the PDF
Image 7
CC-BY · From the PDF
Figure 5 Mean absolute error (MAE) for Height Above Nearest Drainage, Teng Vaze Dutta, and Floodwater Depth Estimation Tool for different depth ranges: shallow (depth <2 m), midrange (depth between 2 and 4 m), and deep (depth greater than 4 m) as well as the MAE for all depth.
Figure 5 Mean absolute error (MAE) for Height Above Nearest Drainage, Teng Vaze Dutta, and Floodwater Depth Estimation Tool for different depth ranges: shallow (depth <2 m), midrange (depth between 2 and 4 m), and deep (depth greater than 4 m) as well as the MAE for all depth.CC-BY · From the PDF
Image 9
CC-BY · From the PDF
Figure 6 Mean absolute error for Height Above Nearest Drainage, Teng Vaze Dutta, and Floodwater Depth Estimation Tool in y -axis plotted against return interval of events (column 1) and geomorphology indices (columns 2-5) in x -axis.
Figure 6 Mean absolute error for Height Above Nearest Drainage, Teng Vaze Dutta, and Floodwater Depth Estimation Tool in y -axis plotted against return interval of events (column 1) and geomorphology indices (columns 2-5) in x -axis.CC-BY · From the PDF
Figure 7 Detailed examination of Floodwater Depth Estimation Tool prediction errors. Part (a) shows the error distribution for each of the 26 flood scenes using a Gaussian kernel density estimate (binwidth = 0.2 m). Plot (b) provides confidence intervals around the median of the series combined using bootstrapping. Plot (c) compares Laplace and Cauchy distributions to the combined distribution. Plot (d) compares the quantiles of the Laplace and Cauchy distributions to the observed error distribution.
Figure 7 Detailed examination of Floodwater Depth Estimation Tool prediction errors. Part (a) shows the error distribution for each of the 26 flood scenes using a Gaussian kernel density estimate (binwidth = 0.2 m). Plot (b) provides confidence intervals around the median of the series combined using bootstrapping. Plot (c) compares Laplace and Cauchy distributions to the combined distribution. Plot (d) compares the quantiles of the Laplace and Cauchy distributions to the observed error distribution.CC-BY · From the PDF
Image 12
CC-BY · From the PDF

Verification

Claims the paper makes

0 Valid · 1 Not valid · 29 Not checked · 86 Own results and interpretations

57 more claims

Create a free account to see all 116 claims and to verify the 25 references still open.

See all claims — free account
  • It may state two separate claims as one. 41
  • Probably not a substantial claim. 23

Create a free account to see these claims.

Citations

Cites

Show all 88 references

Research works printed but not linked: 9

  1. 10.1890/1051-0761 · doi:10.1890/1051-0761
  2. Bowler J. M.(1978).Quaternary climate and tectonics in the evolution of the Riverine Plain southeastern Australia.Landform evolution in Australasia(pp.70–112).
  3. CSIRO, 2008, Water availability in the Lachlan, 133
  4. Davies P. E., 2012, Sustainable rivers audit 2. The ecological health of rivers in the Murray–Darling Basin at the end of the millennium drought (2008–2010), 100
  5. Mackay C. Suter S. Albert N. Morton S. &Yamagata K.(2015).Large scale flexible mesh 2D modelling of the Lower Namoi Valley. InFloodplain Management Association National Conference; Floodplain Management Australia: Brisbane Australia(pp.1–14).
  6. Montazeri M., 2020, Production of 80,000 ML/day flood inundation map for the South Australian section of River Murray
  7. NSW Office of Environment and Heritage, 2017, Lower Namoi Valley floodplain model development, 51
  8. SA Department of Environment and Water, 2020, L3‐L6 calibration report, 30
  9. Teng J. Vaze J. &Dutta D.(2013).Simplified methodology for floodplain inundation modelling using LiDAR DEM. InE.Boegh E.Blyth D. M.Hannah H.Hisdal H.Kunstmann B.Su et al. (Eds.) Climate and land surface changes in hydrology.IAHS Red Book(pp.198–204). Retrieved fromhttp://hdl.handle.net/102.100.100/97033?index=1

Metadata

Authors
Teng, J., Penton, D. J., Ticehurst, C., Sengupta, A., Freebairn, A., Marvanek, S., Vaze, J., Gibbs, M., Streeton, N., Karim, F., & Morton, S.
Affiliations
  • CSIRO Land and Water (AU)
  • NSW Department of Planning and Environment (AU)
Corresponding author
Teng, J.
Venue
Water Resources Research
Published
2022-11
Volume
58
Issue
11
ISSN
  • 0043-1397 (ISSN-L)
  • 1944-7973
OpenAlex
W4307138044
Cited by
42
References
88
Citations by year
2026: 14 · 2025: 16 · 2024: 6 · 2023: 6
FWCI
1.96
Type
journal article
Language
en
Open access
Locations
Water Resources Research (publishedVersion, cc-by)
Funders
  • Australian Government
  • Commonwealth Scientific and Industrial Research Organisation
Topics
  • 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
  • Physical Sciences › Environmental Science › Environmental Engineering › Soil Moisture and Remote Sensing
Keywords
semi-arid region, Height Above Nearest Drainage, digital elevation model, LiDAR DEM, model accuracy, flood magnitude, eastern Australia, flood risk assessment, MERIT DEM, FABDEM

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.

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).

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A Comprehensive Assessment of Floodwater Depth Estimation Models in Semiarid Regions. https://app.cochat.ai/papers/10.1029/2022wr032031
The paper
Teng, J., Penton, D. J., Ticehurst, C., Sengupta, A., Freebairn, A., Marvanek, S., Vaze, J., Gibbs, M., Streeton, N., Karim, F., & Morton, S. (2022). A Comprehensive Assessment of Floodwater Depth Estimation Models in Semiarid Regions. Water Resources Research. https://doi.org/10.1029/2022wr032031

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.