the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Mar 2026
Integrating Satellite Remote Sensing and Local Knowledge to Decipher Two Decades of Coastal Change in the Ensenada de La Paz, Mexico
Abstract. This study presents a comprehensive, two-decade (2005–2025) assessment of coastal morphodynamics in the Ensenada de La Paz, Mexico, by synergistically integrating satellite remote sensing with local knowledge. We employed a multi-sensor approach (Landsat 7 ETM+, Sentinel-2 MSI, and very-high-resolution imagery) to quantify spatiotemporal changes in bathymetry, shoreline position, and turbidity patterns. The analysis reveals a persistent trend of bay-wide shallowing (with an average depth reduction of 0.10 m per five-year period), significant net coastal erosion (21.34 ha), and increased nearshore turbidity, particularly adjacent to urban areas. Hurricanes Newton (2016) and Lorena (2019) triggered distinct, spatially heterogeneous geomorphic responses, driven by differences in rainfall distribution and fluvial sediment inputs. Crucially, structured engagement with local fishers, aquaculturists, and coastal residents provided essential ground-truthing and causal explanations for the remotely sensed patterns, identifying anthropogenic pressures such as illegal fishing, vessel traffic, waste discharge, and proposed infrastructure as key drivers. This integrated framework not only validates local observations with quantitative evidence but also bridges the gap between large-scale change detection and process-based understanding. The findings underscore the system’s vulnerability and provide a robust, evidence-based foundation for participatory coastal management and climate adaptation strategies in data-scarce regions.
- Preprint
(2533 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 07 May 2026)
- CC1: 'Comment on egusphere-2026-983', Luiz Felipe Faria de Sousa, 16 Apr 2026 reply
-
RC1: 'Referee Comment on egusphere-2026-983', Luiz Felipe Faria de Sousa, 17 Apr 2026
reply
The work is very well structured, with a comprehensive methodological approach to answer the questions the authors aim to address. The study addresses sensitive topics that directly deals with anthropogenic impacts on water quality and the side effects on populations that depend on fishing and the recreational use of the area. Studies like this bring society even closer to science and public opinion is an important tool for direct comprehension the magnitude of evaluated events. Thus, I believe it should be published. However, the study can be improved in some aspects related to remote sensing and although the evidence of increased turbidity is notable, the absence of in situ data implies a need for more robust statistical approach to better support the interpretation of results and the recommendations to governmental agencies.
Specific comments:
- The study area map is very well done. However, the upper-left figure could be enlarged to show the entire map of Mexico. Although its title already suggests the country evaluated, a large image helps to better visualize the continent. Additionally, changing the map coordinates to degrees, minutes, and seconds would further facilitate understanding the latitude at which the study region is located, which is important from a geographical perspective.
- Your site description is very well condensed and provides the necessary information. However, adding a bit more information could help you further discuss turbidity dynamics in your channel, such as: (1) Tides are important for water circulation, but what is the tidal range during spring and neap tides in their flood and ebb phases? (2) You mention rainy or extreme periods but I believe it would be important to include information on average rainfall in region, and if possible, temporal information (you could include this in a table in the supplementary material). These data are crucial to support your hypothesis of coastal changes.
- You are using remote sensing tools with excellent spatial visualization capability; however, a fundamental issue concerns me in your work. I would like to better understand the reason for using ETM+ (Landsat-7) data, which has a mechanical failure problem in a component called the Scan Line Corrector (SLC) and consequently around 22% of the scene is compromised by non-data (fill-values). I strongly suggest replace ETM+ (Landsat-7) scenes with TM (Landsat-5), which provides imagery from the 1980s until 2011 (both have the same spatial, temporal and radiometric resolution), and using OLI (Landsat-8, same spatial and temporal resolution but better radiometric resolution), which has been operational since 2013—you can use it until 2015 before switching to Sentinel-2 for subsequent years. It would also be interesting to extend the time series to four decades.
- The authors use different sensors such as Landsat-7 and Sentinel-2. Although the use of a global algorithm is a positive aspect in this study, problems may still exist because the sensors have different spectral response functions. This means that the spectral bandwidth over which upwelling radiation is measured is not the same, and consequently, if Sentinel-2 and Landsat-7 images were acquired on the same day, turbidity results would be inconsistent between sensors. In practical terms, turbidity may be overestimated or underestimated by one sensor relative to the other, leading to systematic errors. In this case, I strongly suggest applying a technique known as spectral harmonization (e.g., Scheffler et al., 2020: https://doi.org/10.1016/j.rse.2020.111723).
- During extreme events, turbidity can increase substantially and in such cases the red band would not be a good predictor. This is explained by the higher intensity of multiple scattering by particles (the absorption by “pure” water becomes less significant), reaching its asymptotic limit (saturation) more quickly than longer wavelengths such as the near-infrared (NIR). In this case, it would be interesting to include the Nechad et al. (2016) algorithm for NIR band in your turbidity analyses for extreme events. It would be even more interesting to apply a hybrid approach (using reflectance thresholds to switch between the red and NIR algorithms as discussed by Dogliotti et al., 2015).
- Siltation is undeniable evidence for increased turbidity due to the greater potential to keep particles suspended. However, to what extent is the increase in turbidity truly suspended material versus bottom signal and/or adjacency effects? Since siltation brings the bottom closer to the surface, comparing a single image from 2005 with another from 2025 may introduce confirmation bias. Although this is somewhat obvious evidence, it would be useful to present some statistical results to strengthen your recommendation for installing an in situ station. I suggest that the authors select some pixels in strategic regions using the entire available time series (Landsat 5, 7, 8, and 9 + Sentinel-2), properly masked for clouds, uncorrectable glint, and land features (see the MNDWI index at https://www.tandfonline.com/doi/full/10.1080/01431160600589179). Note that here Landsat-7 issues are less problematic because you can work with individual pixels rather than spatial interpretations. With a time series, the authors can apply the Mann-Kendall method to assess whether the turbidity increase is statistically significant.
Other comments:
Line 71 - 74: Acolite is a processor that includes different atmospheric correction methods, such as dark spectrum function, exponential and RadCOR. By default, the atmospheric correction is the dark spectrum function. You need to include this as information in the text.Line 80 - 105: The relative depth method is interesting but still depends on water signal. Therefore, it is necessary to include some information regarding the uncertainty of this model. Does it work up to 5 to 10 m? In this sense, it is necessary to include a brief discussion of the limitations of this type of approach, considering each optically active constituent (OAC) (under which optical conditions was it developed?), since each OAC has specific absorption and scattering processes.
Line 111 - 113: The same applies to the Nechad et al. (2016) algorithm. Include the model uncertainties (results from the validations presented by the authors in the original paper). In the discussion, present a brief limitation of the algorithm, such as not being parameterized for the study area, but especially its physical assumptions. For example, in its formulation, the algorithm assumes that light absorption by chromophoric dissolved organic matter is spatially invariant, which is not true, since in your study area there may be different types of CDOM, each with distinct light absorption characteristics.
Citation: https://doi.org/10.5194/egusphere-2026-983-RC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 118 | 49 | 13 | 180 | 13 | 19 |
- HTML: 118
- PDF: 49
- XML: 13
- Total: 180
- BibTeX: 13
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Faria de sousa, Luiz Felipe Machado:
Decision: Major reviews
The work is very well structured, with a comprehensive methodological approach to answer the questions the authors aim to address. The study addresses sensitive topics that directly deals with anthropogenic impacts on water quality and the side effects on populations that depend on fishing and the recreational use of the area. Studies like this bring society even closer to science and public opinion is an important tool for direct comprehension the magnitude of evaluated events. Thus, I believe it should be published. However, the study can be improved in some aspects related to remote sensing and although the evidence of increased turbidity is notable, the absence of in situ data implies a need for more robust statistical approach to better support the interpretation of results and the recommendations to governmental agencies.
Specific comments: