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

The DOI matches its Crossref record; 56 of the 56 references it cites were found in a registry or an index; none of the 16 sources its claims cite is checked for support yet.

Space‐Borne Cloud‐Native Satellite‐Derived Bathymetry (SDB) Models Using ICESat‐2 And Sentinel‐2

Open access
Thomas, N., Pertiwi, A. P., Traganos, D., Lagomasino, D., Poursanidis, D., Moreno, S., Fatoyinbo, L. · 2021 · Geophysical Research Letters

Published version of a preprint

Abstract

Shallow nearshore coastal waters provide a wealth of societal, economic, and ecosystem services, yet their topographic structure is poorly mapped due to a reliance upon expensive and time intensive methods. Space‐borne bathymetric mapping has helped address these issues, but has remained largely dependent upon in situ measurements. Here we fuse ICESat‐2 lidar data with Sentinel‐2 optical imagery, within the Google Earth Engine cloud platform, to create openly available spatially continuous high‐resolution bathymetric maps at regional‐to‐national scales in Florida, Crete and Bermuda. ICESat‐2 bathymetric classified photons are used to train three Satellite Derived Bathymetry (SDB) methods, including Lyzenga, Stumpf, and Support Vector Regression algorithms. For each study site the Lyzenga algorithm yielded the lowest RMSE (approx. 10%–15%) when compared with validation data. We demonstrate a means of using ICESat‐2 for both model calibration and validation, thus cementing a pathway for fully space‐borne estimates of nearshore bathymetry in shallow, clear water environments.

Overview

CoChat Summary

TL;DR

The paper fuses ICESat-2 lidar and Sentinel-2 imagery in Google Earth Engine to produce fully space-borne satellite-derived bathymetry maps in Bermuda, Biscayne Bay, and Crete. Across three SDB methods, cluster-based Lyzenga (CBL) gave the lowest errors, with RMSE 2.62 m, 0.83 m, and 2.19 m respectively.

Key findings

  • CBL produced the most reliable SDB estimates at all sites: RMSE 2.62 m (Bermuda), 0.83 m (Biscayne Bay), and 2.19 m (Crete); MAE 2.00 m, 0.65 m, and 2.02 m; R² 0.68, 0.79, and 0.83; maximum modeled depths 26 m, 5 m, and 22 m.
  • CBL RMSEs were lower than or close to 10% of maximum depth for Bermuda and Crete, but close to 17% for Biscayne Bay; SVR tended to underestimate the range of depths although its errors did not differ much from CBL.
  • L2A and L1C errors were comparable, indicating that computationally expensive atmospheric correction is not needed for accurate SDB retrieval in shallow coastal environments.
  • ICESat-2 was used for both model calibration and validation, supporting a pathway to fully space-borne nearshore bathymetry; the SDB models captured bathymetric relief and rugosity in shallow and coral reef areas in more detail than the NOAA DEM.

Methods

  • A cloud-native Google Earth Engine workflow combined multi-temporal Sentinel-2 A/B data (B1–B4, 20th-percentile mosaic after cloud masking) with ICESat-2 ATL03 photons corrected for refraction, using high-power beams and “Land Confidence” level 4 photons aggregated to a 10-m Sentinel-2 grid.
  • Three SDB models were trained: CBL (cluster-based regression with Lyzenga log-linear regression using B1–B4), CBS (cluster-based regression with Stumpf log-ratio using B3/B2), and SVR (epsilon-SVR with linear kernel on the first four log-transformed Sentinel-2 bands); k-means clustering with five clusters was used for Bermuda and Biscayne Bay, but not for Crete due to its predominantly homogeneous sandy seabed.
  • Accuracy assessment used an 80/20 training/validation split: Bermuda used 5,173 ICESat-2 training points and 1,293 ICESat-2 validation points; Biscayne Bay used 34,342 ICESat-2 training points and 8,585 NOAA DEM validation points; Crete used 133 ICESat-2 training points and 85 in-situ singlebeam validation points; metrics were RMSE, MAE, R², and standard deviation.

Limitations

  • Stated: manual selection of ICESat-2 bathymetric photons is a potential source of user-interpretation error, especially where ICESat-2 served as both training and validation data (Bermuda and Crete); tide-level mismatch between Sentinel-2 20th-percentile composites and ICESat-2 acquisition dates is non-negligible but assumed small for the microtidal (<1 m) study sites.
  • Stated: Biscayne Bay ICESat-2 could not detect the bottom of dredged channels due to sediment/turbidity or limited photon penetration; Crete SDB underpredicted depths greater than 15 m where water is optically deep and the relationship becomes non-linear; Bermuda NOAA DEM validation was poor (r² = 0.14) due to low temporal/spatial resolution and interpolation.
  • Evident: validation data quality and independence vary by site (e.g., Crete validation used only 85 in-situ points; Bermuda validation used ICESat-2, though from different passes/days), and the approach is limited to shallow, clear, optically shallow water environments.

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

Integrity signals

Registry

Crossref · 10.1029/2020gl092170

Cites

56 of 56 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 (a, b & c) ICESat-2 depth data points for Bermuda, Biscayne Bay and Crete; (d, e & f) ICESat-2 depth data points with 'Land Confidence' level 4 binned to Sentinel-2 10-m resolution for Bermuda (SDB training, green and validation, blue), Biscayne Bay and Crete; (g, h & i) ICESat-2 photons and selected bathymetric profile for a single laser transect in Bermuda, Biscayne Bay and Crete.
Figure 1 (a, b & c) ICESat-2 depth data points for Bermuda, Biscayne Bay and Crete; (d, e & f) ICESat-2 depth data points with 'Land Confidence' level 4 binned to Sentinel-2 10-m resolution for Bermuda (SDB training, green and validation, blue), Biscayne Bay and Crete; (g, h & i) ICESat-2 photons and selected bathymetric profile for a single laser transect in Bermuda, Biscayne Bay and Crete.CC-BY · From the PDF
Image 2
CC-BY · From the PDF
Figure 2 (a, b, c) Sentinel-2 RGB synthesis. (a) Bermuda: 53 L2A Surface Reflectance tiles, 597 km 2 (March 28, 2017 - April 20, 2020); (b) Biscayne Bay: 583 L1C Top-Of-Atmosphere tiles, 689 km 2 (January 1, 2015 - December 31,2019); (c) Crete: L1C 403L1C Top-Of-Atmosphere, 61 km 2 (January 1, 2015 - December 31,2019). (d, e, f) CBL Bathymetry SDB at Bermuda, Biscayne Bay and Crete. (g, h, i) NOAA DEM at Bermuda and Biscayne Bay and EMODnet at Crete, j, k, l) SDB-ICESat-2 depth comparison at Bermuda, Biscayne Bay and Crete.
Figure 2 (a, b, c) Sentinel-2 RGB synthesis. (a) Bermuda: 53 L2A Surface Reflectance tiles, 597 km 2 (March 28, 2017 - April 20, 2020); (b) Biscayne Bay: 583 L1C Top-Of-Atmosphere tiles, 689 km 2 (January 1, 2015 - December 31,2019); (c) Crete: L1C 403L1C Top-Of-Atmosphere, 61 km 2 (January 1, 2015 - December 31,2019). (d, e, f) CBL Bathymetry SDB at Bermuda, Biscayne Bay and Crete. (g, h, i) NOAA DEM at Bermuda and Biscayne Bay and EMODnet at Crete, j, k, l) SDB-ICESat-2 depth comparison at Bermuda, Biscayne Bay and Crete.CC-BY · From the PDF
Image 4
CC-BY · From the PDF

Verification

Claims the paper makes

0 Valid · 1 Not valid · 13 Not checked · 54 Own results and interpretations

29 more claims

Create a free account to see all 68 claims and to verify the 13 references still open.

See all claims — free account
  • It may state two separate claims as one. 23
  • Probably not a substantial claim. 13

Create a free account to see these claims.

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.

3 of 3

Citations

Cites

Show all 56 references

Research works printed but not linked: 5

  1. Daly C. J. Baba W. Bergsma E. Almar R. &Garlan T.(2020).The New Era of Regional Coastal Bathymetry fromSpace: A Showcase for West Africa using Sentinel‐2Imagery. Earth ArXiv https://doi.org/10.31223/osf.io/f37rv · doi:10.31223/osf.io/f37rv
  2. 10.1016/fj.marpol.2010.02.001 · doi:10.1016/fj.marpol.2010.02.001
  3. Marks K., 2019, The IHO‐IOC GEBCO cook book
  4. National Centers for Environmental Information, 2018, NOAA NOS Estuarine Bathymetry ‐ Biscayne Bay (S200)
  5. 10.3389/fmars.2019.00283/full · doi:10.3389/fmars.2019.00283/full

Cited by

Metadata

Authors
Thomas, N., Pertiwi, A. P., Traganos, D., Lagomasino, D., Poursanidis, D., Moreno, S., & Fatoyinbo, L.
Affiliations
  • Goddard Space Flight Center (US)
  • Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR) (DE)
  • East Carolina University (US)
  • FORTH Institute of Applied and Computational Mathematics (GR)
  • NASA Goddard Space Flight Center Biospheric Sciences Laboratory Greenbelt MD USA
Corresponding author
Thomas, N.
Venue
Geophysical Research Letters
Published
2021-03-22
Volume
48
Issue
6
ISSN
  • 0094-8276 (ISSN-L)
  • 1944-8007
OpenAlex
W3090911235
Cited by
154
References
56
Citations by year
2026: 25 · 2025: 30 · 2024: 26 · 2023: 32 · 2022: 23 · 2021: 18
FWCI
16.79
Type
journal article
Language
en
Open access
Locations
  • Geophysical Research Letters (publishedVersion, cc-by)
  • Edge Hill University Research Information Repository (Edge Hill University) (publishedVersion, cc-by)
  • elib (German Aerospace Center) (acceptedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • The Scholarship East Carolina University's Institutional Repository (East Carolina University) (submittedVersion)
Funders
National Aeronautics and Space Administration (80NSSC20K0968, NNX17AE79A)
Topics
  • Life Sciences › Agricultural and Biological Sciences › Forestry › Remote Sensing and LiDAR Applications
  • Physical Sciences › Earth and Planetary Sciences › Earth-Surface Processes › Coastal and Marine Dynamics
  • Physical Sciences › Earth and Planetary Sciences › Oceanography › Underwater Acoustics Research
Keywords
satellite-derived bathymetry, ICESat-2, Sentinel-2, Lyzenga algorithm, support vector regression, Florida, Crete, Bermuda

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 2026-11-18, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-18 (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

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.