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

The DOI matches its Crossref record; 87 of the 88 references it cites were found in a registry or an index; none of the 91 sources its claims cite is checked for support yet.

Assessment of River Planform Dynamics in the Amazon Basin Using Sentinel-1 SAR Data (2017–2025)

Open access
Rijt, Ivar van, Balling, Johannes, Reiche, Johannes · 2026 · Remote Sensing

Abstract

The Amazon Basin and its rivers play a vital role in regional biodiversity, the carbon cycle, and socio-economic security. Through erosion and deposition, river planforms change over time, affecting local infrastructure, food security, and changes to ecosystems. Long-term monitoring is essential for observing these dynamics. Synthetic Aperture Radar (SAR) provides a method to consistently map river planform dynamics across large areas because it is largely independent of atmospheric conditions. This study presents an approach for deriving river planform metrics across the entire Amazon Basin using Sentinel-1 C-band SAR data. This approach followed three main steps: water mask generation, validation of the data and river metrics extraction. Sentinel-1 imagery from 2017 to 2025 was composited into quarterly mean images, after which Otsu thresholding was applied to derive water classifications. Additional post-processing steps were applied to reduce terrain- and seasonal effects. The final water masks were divided into water-change classes, validated using stratified sampling and achieved an overall accuracy of 98.5%. Quarterly river planform metrics, including sinuosity, mean channel width and migration rate, were derived using channel centerline extraction, but due to a lack of in situ validation data the river metric values have not been validated. The resulting time series provide insights into how river planform changes across all Amazon sub-basins from 2017 to 2025 can be monitored using SAR-based methods. The results reveal spatial differences in river dynamics between tributaries, mostly depending on flow pattern, up- or downstream path and location in the upper, middle or lower Amazon Basin. These findings demonstrate the potential of SAR time series for monitoring large-scale river planform dynamics.

Overview

CoChat Summary

TL;DR

This study develops and evaluates a basin-wide framework for extracting river planform metrics (mean channel width, sinuosity, migration rate, and water-change classes) across the Amazon Basin from Sentinel-1 SAR time series (2017–2025). It shows that SAR-derived quarterly water masks can support consistent large-scale geomorphological monitoring, though water gain/loss classes and metric validation remain limited.

Key findings

  • The final dataset comprises 36 quarterly binary water masks covering the full Amazon Basin (2017–2025) with an overall accuracy of 98.48%; class-wise, non-water reached UA 99% / PA 99.8% and stable water UA 90% / PA 78.3%, while water gain (UA 38%, PA 42.45%) and water loss (UA 70%, PA 25.03%) were much lower.
  • Mean channel width ranged from 96.04 m (Javari) to 2387.47 m (Amazon floodplain), with Tapajós at 1978.84 m and Rio Araua at 1786.97 m; width appeared to scale with sub-basin surface area.
  • Migration rate was the most spatially heterogeneous metric, ranging from 2.10 to 293.94 m/quarter with no clear up- or downstream gradient; the upper Ucayali was highest (293.94 m/q), Xingu second (96.81 m/q), and the Amazon floodplain moderate (67.02 m/q).
  • Sinuosity ranged from 1.42 to 2.67, distinguishing anastomosing from meandering planforms: Amazon floodplain (1.52), Madeira (1.52) and Negro (1.71) were low, while Purus (2.67), Javari (2.64) and Juruá (2.58) were high. The 2023 El Niño event produced the minimum recorded stable water area in the Amazon floodplain.

Methods

  • Sentinel-1 GRD imagery (IW mode, VH polarization, 10 m pixel spacing) from 1 January 2017 to 31 December 2025 was processed in Google Earth Engine; 22,133 ascending and 65,121 descending images were collected and composited into 3-month intervals (36 composites total).
  • Water was classified using Otsu thresholding (threshold −17.4 dB) after comparison with a histogram-based threshold (−24.6 dB), followed by terrain (elevation >1500 m, slope >20°) and ESA WorldCover landcover filtering, a 1024-pixel connectivity filter, a temporal majority filter, and a 5×5 binary-closing gap-fill.
  • Validation used stratified random sampling (50 points per class per map across 9 water-change maps, 1800 points total) against PlanetScope quarterly mosaics, with area-weighted accuracy assessment (Olofsson et al.) in R; centerlines were extracted via skeletonization and a width-weighted Dijkstra shortest-path algorithm with Savitzky-Golay smoothing, and metrics computed per quarter across 29 sub-basins.

Limitations

  • Stated: the failure of Sentinel-1B in late 2021 caused major data gaps (Japurá-Caquetá, Madeira, Negro), reducing measurement consistency until Sentinel-1C launched in early 2025; water gain/loss classes are sensitive to mixed edge pixels and seasonal effects; Otsu assumes bimodality while Amazon backscatter is unimodal; temporal compositing can obscure short-term flood processes; and there is no ground-truth or independent reference data to validate the derived metrics, so literature comparison was used as a "suboptimal" proxy.
  • Stated: C-band SAR has limited canopy penetration, so narrow channels and floodplains under closed canopy may be systematically under-detected; remaining water-mask gaps produce missing (NaN) metric values; and sub-basin mean reporting risks missing local changes such as cutoffs.
  • Evident from the text: the unusually high Ucayali migration rate (293.94 m/q) is attributed by the authors to noise from classification errors, smoothing, and sudden events rather than true migration, and the paper notes the uncertainty through the full processing pipeline is "complex and difficult to quantify."

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

Integrity signals

Registry

Crossref · 10.3390/rs18132075

Cites

87 of 87 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Top 15 most used planform river dynamics sorted by count based on the author's semisystematic literature review of 60 peer-reviewed sources.
Figure 1 Top 15 most used planform river dynamics sorted by count based on the author's semisystematic literature review of 60 peer-reviewed sources.CC-BY · From the PDF
Figure 2 Percentage of most used remote sensing and other sensors in literature based on 60 reviewed sources. Spectral is the most used remote sensing sensor with 66.7%. The 'other sensors' category consists of non-remote sensing methods such as tacheometry, photography, water level gauging, sonar, kinematic DGPS and hydrometers.
Figure 2 Percentage of most used remote sensing and other sensors in literature based on 60 reviewed sources. Spectral is the most used remote sensing sensor with 66.7%. The 'other sensors' category consists of non-remote sensing methods such as tacheometry, photography, water level gauging, sonar, kinematic DGPS and hydrometers.CC-BY · From the PDF
Figure 3 ( A ) Study area reference within the continent of South America. ( B ) Twenty nine BL2 hydrological Amazonian sub-basins, based on Venticinque et al. [7]. ( C ) Geotectonic units of the Amazon Basin: Acre, Solimões, Amazonas and Marajó basins, modified from Rozo et al. [16].
Figure 3 ( A ) Study area reference within the continent of South America. ( B ) Twenty nine BL2 hydrological Amazonian sub-basins, based on Venticinque et al. [7]. ( C ) Geotectonic units of the Amazon Basin: Acre, Solimões, Amazonas and Marajó basins, modified from Rozo et al. [16].CC-BY · From the PDF
Image 4
CC-BY · From the PDF
Figure 5 Flowchart of the methodology, including data input, pre-processing in GEE, water classification in GEE, post-processing in Python (v3.13) and validation in QGIS (v3.40.12) and R (v4.4.2), for a larger version of this figure see Supplement S1.
Figure 5 Flowchart of the methodology, including data input, pre-processing in GEE, water classification in GEE, post-processing in Python (v3.13) and validation in QGIS (v3.40.12) and R (v4.4.2), for a larger version of this figure see Supplement S1.CC-BY · From the PDF
Figure 6 Graph showing the distribution of VH backscatter values (in dB) and the fixed histogram threshold (green) and Otsu threshold (red).
Figure 6 Graph showing the distribution of VH backscatter values (in dB) and the fixed histogram threshold (green) and Otsu threshold (red).CC-BY · From the PDF
Figure 7 ( A1 ) PlanetScope satellite imagery of a turbulent water surface and ( B1 ) of an anastomosing river pattern. ( A2 , B2 ) Water classification using the fixed histogram threshold and ( A3 , B3 ) water classification using Otsu. ( A2 , B2 ) More gaps and inconsistencies in water classification than ( A3 , B3 ).
Figure 7 ( A1 ) PlanetScope satellite imagery of a turbulent water surface and ( B1 ) of an anastomosing river pattern. ( A2 , B2 ) Water classification using the fixed histogram threshold and ( A3 , B3 ) water classification using Otsu. ( A2 , B2 ) More gaps and inconsistencies in water classification than ( A3 , B3 ).CC-BY · From the PDF
Figure 8 ( A ) ESA WorldCover 10 m v200 dataset. ( B ) Water classification using the Otsu threshold without implementing a landcover filter, where all blue areas are water and white areas are nonwater. ( C ) Water classification using the Otsu threshold with the landcover-based filter.
Figure 8 ( A ) ESA WorldCover 10 m v200 dataset. ( B ) Water classification using the Otsu threshold without implementing a landcover filter, where all blue areas are water and white areas are nonwater. ( C ) Water classification using the Otsu threshold with the landcover-based filter.CC-BY · From the PDF
Figure 9 ( A ) PlanetScope satellite imagery of a section of the Rio Branco in Boa Vista where a bridge and a patch of bare soil are visible. ( B ) Raw water mask as exported from GEE. ( C ) Water mask after applying the majority filter. ( D ) Water mask after applying the gap-filling filter.
Figure 9 ( A ) PlanetScope satellite imagery of a section of the Rio Branco in Boa Vista where a bridge and a patch of bare soil are visible. ( B ) Raw water mask as exported from GEE. ( C ) Water mask after applying the majority filter. ( D ) Water mask after applying the gap-filling filter.CC-BY · From the PDF
Figure 10 Flowchart showing planform river metric extraction steps, such as preparation, centerline extraction and metric calculation, for a larger version of this figure see Supplement S2.
Figure 10 Flowchart showing planform river metric extraction steps, such as preparation, centerline extraction and metric calculation, for a larger version of this figure see Supplement S2.CC-BY · From the PDF
Figure 11 Boxplots showing all water-change classes and their respective User and Producer accuracies across all periods (nine per class).
Figure 11 Boxplots showing all water-change classes and their respective User and Producer accuracies across all periods (nine per class).CC-BY · From the PDF
Figure 12 All generated centerlines of each Amazon tributary mainstem of 2019 Q2 and their respective basins.
Figure 12 All generated centerlines of each Amazon tributary mainstem of 2019 Q2 and their respective basins.CC-BY · From the PDF
Image 13
CC-BY · From the PDF
Figure 14 Mean of the six river metrics for each Amazon sub-basin: ( A ) Mean channel width. ( B ) Mean migration rate. ( C ) Mean sinuosity. ( D ) Mean stable water area. ( E ) Mean water gain area. ( F ) Mean water loss area.
Figure 14 Mean of the six river metrics for each Amazon sub-basin: ( A ) Mean channel width. ( B ) Mean migration rate. ( C ) Mean sinuosity. ( D ) Mean stable water area. ( E ) Mean water gain area. ( F ) Mean water loss area.CC-BY · From the PDF
Figure A1. Class-wise User and Producer accuracy for each validated period.
Figure A1. Class-wise User and Producer accuracy for each validated period.CC-BY · From the PDF
Figure A2. Confusion Matrix counts: 2017 Q2 to 2017 Q3.
Figure A2. Confusion Matrix counts: 2017 Q2 to 2017 Q3.CC-BY · From the PDF
Figure A3. Confusion Matrix counts: 2018 Q2 to 2018 Q3.
Figure A3. Confusion Matrix counts: 2018 Q2 to 2018 Q3.CC-BY · From the PDF
Figure A4. Confusion Matrix counts: 2019 Q3 to 2019 Q4.
Figure A4. Confusion Matrix counts: 2019 Q3 to 2019 Q4.CC-BY · From the PDF
Figure A5. Confusion Matrix counts: 2020 Q1 to 2020 Q2.
Figure A5. Confusion Matrix counts: 2020 Q1 to 2020 Q2.CC-BY · From the PDF
Figure A6. Confusion Matrix counts: 2021 Q3 to 2021 Q4.
Figure A6. Confusion Matrix counts: 2021 Q3 to 2021 Q4.CC-BY · From the PDF
Figure A7. Confusion Matrix counts: 2022 Q2 to 2022 Q3.
Figure A7. Confusion Matrix counts: 2022 Q2 to 2022 Q3.CC-BY · From the PDF
Figure A8. Confusion Matrix counts: 2023 Q3 to 2023 Q4.
Figure A8. Confusion Matrix counts: 2023 Q3 to 2023 Q4.CC-BY · From the PDF
Figure A9. Confusion Matrix counts: 2024 Q1 to 2024 Q2.
Figure A9. Confusion Matrix counts: 2024 Q1 to 2024 Q2.CC-BY · From the PDF
Figure A10. Confusion Matrix counts: 2025 Q2 to 2025 Q3.
Figure A10. Confusion Matrix counts: 2025 Q2 to 2025 Q3.CC-BY · From the PDF

Verification

Claims the paper makes

0 Valid · 0 Not valid · 59 Not checked · 151 Own results and interpretations

69 more claims

Create a free account to see all 210 claims and to verify the 91 references still open.

See all claims — free account
  • It may state two separate claims as one. 74
  • Probably not a substantial claim. 70
  • The quote only partly states this claim. 7

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.

5 of 6 · the rest aren’t loaded

1 more finding

Create a free account to see every finding of this check.

See all findings — free account

Citations

Cites

Research works printed but not linked: 2

  1. Christopherson, 2015, River Systems, Geosystems, An Introduction to Physical Geography, Volume 9, 448
  2. Felzenszwalb, 2012, Distance Transforms of Sampled Functions, Theory Comput., 8, 415, 10.4086/toc.2012.v008a019 · doi:10.4086/toc.2012.v008a019

Metadata

Authors
Rijt, I. V., Balling, J., & Reiche, J.
Affiliations
  • Laboratory of Geo-Information Science and Remote Sensing, Wageningen University & Research, Droevendaalsesteeg 3, 6708 PB Wageningen, The Netherlands
  • Wageningen University & Research (NL)
Corresponding author
Reiche, Johannes
Venue
Remote Sensing
Publisher
MDPI AG
Published
2026-06-24
Volume
18
Issue
13
Pages
2075
ISSN
2072-4292 (ISSN-L)
OpenAlex
W7165745692
Cited by
1
References
88
Citations by year
2026: 1
FWCI
2.29
Type
journal article
Language
en
Open access
Locations
  • Remote Sensing (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Wageningen University & Research Publications (submittedVersion, cc-by)
Funders
Wageningen University and Research
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 › Earth and Planetary Sciences › Geophysics › Synthetic Aperture Radar (SAR) Applications and Techniques
Keywords
Amazon basin rivers, Sentinel-1 SAR, Otsu thresholding, river sinuosity, channel width, tributaries

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.