Remote sensing as a tool for assessing landslide impacts on coastal waters at Porchat Island (SE Brazil)

Remote sensing as a tool for assessing landslide
impacts on coastal waters at Porchat Island (SE
Brazil)
Thainá de Oliveira Altafini 
Universidade Federal de São Paulo
Beatriz Silva Fernandes 
Universidade Federal de São Paulo
Breylla Campos Carvalho 
Universidade Federal de São Paulo
Research Article
Keywords: Remote sensing, coastal management, turbidity, suspended matter, chlorophyll-a
Posted Date: October 23rd, 2026
DOI: https://doi.org/10.21203/rs.3.rs-11065538/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License.  
Read Full License
Additional Declarations: No competing interests reported.
Page 1/24

Abstract
On April 29, 2025, a water pipeline rupture on Porchat Island (São Vicente, southeastern Brazil) triggered
a mass movement that released sediment into adjacent coastal waters. Effects of urban infrastructure
failures on coastal water quality are rarely documented with imagery of adequate spatial and temporal
resolution. We used Sentinel-2 and PlanetScope imagery, processed in ACOLITE, to estimate turbidity,
suspended matter concentration (SMC), chlorophyll-a, and Secchi disk depth before and after the
rupture, and related the sediment plume to meteoceanographic conditions. In Sentinel-2 images, mean
SMC in the waters adjacent to the island increased from 11.15 to 21.09 g/m³ and mean Secchi depth
decreased from 2.01 to 1.12 m, with smaller changes in Santos and São Vicente bays. PlanetScope
images showed a similar pattern near the island, with values of similar magnitude. Chlorophyll-a varied
among dates without a clear response. As no in situ data were available, results are interpreted as
relative changes. PlanetScope's daily revisit showed the sediment plume shrinking from 7.50 km² on
April 30 to 2.02 km² on May 2 and dispersing toward central Santos Bay under waves from the southeast
and south quadrants and weak winds. This case shows that an infrastructure failure can cause marked
but apparently temporary changes in the optical properties of coastal waters, and indicates that high-
frequency, high-resolution remote sensing is useful for rapid impact assessment and for supporting
coastal monitoring and management.
1 Introduction
Coastal zones play a fundamental role in maintaining ecosystem services, including biodiversity
conservation, carbon sequestration, and natural protection against extreme events such as storms and
coastal flooding (Barbier et al. 2011; Hopkinson et al. 2019; Windham-Myers, Crooks, Troxler 2018). The
maintenance of these environments depends directly on several factors, such as the continuous input of
sediments from rivers, coastal erosion, and slope instability processes, considering that sediment
deficits can compromise the integrity of coastal morphology and reduce the capacity of these systems
to respond to environmental pressures (Coleman et al. 2022; Fagherazzi et al. 2013; Kirwan, Megonigal
2013; Ladd et al. 2019).
Among the sediment sources that influence coastal environments, mass movement processes stand
out, playing an important role in landscape evolution, weathering processes, and sediment dynamics
(Dadson et al. 2003; Neil et al. 2002; Rabalais et al. 2007). These events can release large volumes of
terrigenous material over short periods of time, significantly altering sediment fluxes between
continental and coastal environments. Studies show that episodic events, such as floods, intense
storms, and landslides, can promote significant changes in aquatic ecosystems by increasing sediment
and organic matter input to the coastal zone (Fong, Gaynus, Carpenter 2020; Smith et al. 2021). As a
result, fast changes may occur in the physical, chemical, and biological characteristics of the water
column, including changes in turbidity, transparency, light availability, and suspended matter
concentration (SMC), parameters directly related to water quality and ecosystem functioning (Kobiyama
et al. 2011; Simão et al. 2025).
Page 2/24

The changes are largely governed by the optically active components of coastal waters, namely
dissolved and particulate substances of organic and inorganic origin that interact with solar radiation
through absorption and scattering (Barbosa et al. 2019; Mobley 1994). Among these constituents, total
particulate matter (or total suspended matter), composed of phytoplankton and non-algal inorganic and
organic particles, strongly influences the optical behavior of water, controlling light penetration through
the water column, reducing the availability of photosynthetically active radiation, limiting primary
production, and influencing the functioning of aquatic ecosystems (Mobley 1994, Li et al. 2022; Wang et
al. 2022).
Another important optically active component is chlorophyll-a, the main photosynthetic pigment of
phytoplankton, widely used as an indicator of phytoplankton biomass, primary productivity, and the
eutrophication status of aquatic environments (Boyce, Lewis, Worm 2010; Gohin et al. 2019). In coastal
and estuarine areas, its distribution shows high spatial and temporal variability due to the mixing of
continental and marine waters, nutrient inputs, and local hydro-sedimentary dynamics (Arabi et al. 2020;
Gitelson et al. 2007). Thus, the variables are useful indicators for coastal water quality assessment and
management.
Monitoring these parameters is essential for understanding water dynamics and quality. However,
conventional methods based on field sampling campaigns and laboratory analyses require substantial
financial investment, human resources, and execution time, in addition to presenting limitations
regarding the spatial and temporal coverage of observations (Bende-Michl et al. 2011; Brivio, Giardino,
Zilioli 2001; Schaeffer et al. 2013; Volk, Liersch, Schmidt 2009).
In this context, remote sensing has become established as an important tool for the environmental
monitoring of water bodies. Technological advances in orbital sensors have improved the spatial and
temporal resolution of imagery, providing spatially explicit information that can support rapid impact
assessment, emergency response, and coastal management (Barbosa et al. 2019; Jensen 2014).
Despite the progress observed in the remote estimation of parameters such as turbidity, SMC, and
chlorophyll-a, studies evaluating the response of coastal waters to episodic sediment input events
resulting from mass movements remain scarce (e.g., Irie, Nakagawa, Higashi 2023; Kobiyama et al.
2011; Lin et al. 2011), and even fewer have done so with spatial resolutions compatible with
management needs, such as delimiting the affected area and tracking its persistence. This gap becomes
particularly relevant in urbanized coastal regions, where anthropogenic interventions can trigger abrupt
changes in hydro-sedimentary processes and where information on such events is needed to guide
response, impact assessment, and slope and drainage management.
One such event occurred on April 29, 2025, when the rupture of a water pipeline belonging to the
Companhia de Saneamento Básico do Estado de São Paulo (Sabesp - Basic Sanitation Company of the
State of São Paulo), located on Porchat Island, in the municipality of São Vicente (São Paulo, Brazil),
triggered a mass movement on an adjacent slope. The large volume of water released caused the
displacement of sediment and vegetation (Santa Portal 2025), estimated at approximately 860,000 m³
Page 3/24

(Mário Bueno, personal communication, June 23, 2025), resulting in the transport of this material to the
coastal waters near the island (G1 Santos 2025). Porchat Island is an urbanized area on steep slopes,
classified as a Permanent Preservation Area and subject to civil defense monitoring for geological risk,
which makes the event a representative case of the interaction between urban infrastructure, slope
stability, and coastal water quality (Marques 2015). The event caused a marked visual change in water
color, suggesting substantial changes in SMC and other water quality-related parameters.
Given this context, the present study aimed to quantify turbidity, water transparency, SMC, and
chlorophyll-a concentrations before and after the pipeline rupture that occurred in April 2025, using high
spatial resolution optical imagery, and to delineate the extent and persistence of the sediment plume in
relation to the prevailing meteoceanographic conditions. By doing so, we evaluate the potential of high-
resolution remote sensing as a tool for rapid impact assessment, providing support for environmental
monitoring, coastal planning and management, and the assessment of the effects of urban
infrastructure failures on coastal marine environments.
2 Materials and methods
2.1 Study area
Porchat Island is located in the municipality of São Vicente (Fig. 1), on the central coast of the state of
São Paulo, forming part of the Santos Estuarine System (SES), a hydrological network that discharges
into the Santos and São Vicente channels. The waters adjacent to the island connect Santos Bay (SB)
and São Vicente Bay (SVB). The oceanographic dynamics of the bays are marked by high hydrodynamic
complexity, governed mainly by a mixed microtidal regime (Seiler 2019).
From a geomorphological standpoint, the area is part of the Residual Hills compartment, situated at the
base of the Serra do Mar, with elevations reaching approximately 200 meters. The local geology is
predominantly composed of Precambrian rocks, such as migmatites, gneisses, and granites, resulting
from the tilting of the Serra do Mar toward the ocean. The topography is markedly rugged, with steep
slopes averaging 50° in inclination, and due to the abrupt contact between the hills and the sea, the
island is characterized by rocky shores and cliffs, which prevent the natural formation of beaches along
its immediate perimeter (Marques 2015).
Anthropogenic interference on Porchat Island intensified from the 1950s onward, evolving into
consolidated urbanization with multi-story buildings and disorderly occupation, often on shallow residual
soils, increasing susceptibility to mass movements. Soil sealing caused by paving and construction,
combined with drainage systems that frequently become insufficient or obstructed by debris, overloads
the slopes and favors erosive processes and landslides (Marques 2015).
Environmentally, Porchat Island is classified as a Permanent Preservation Area (APP), protected under
the Forest Code and federal and state legislation to mitigate erosion and preserve Atlantic Forest
biodiversity. However, urban pressure and historical deforestation have generated a state of physical
Page 4/24

fragility, making the area subject to constant monitoring by civil defense preventive plans due to the risk
of geological accidents triggered by intense rainfall events (Marques 2015).
2.2 Meteoceanographic data
The in situ wave data used in this study, significant wave height (Hs) and period (T), were provided by the
Santos Port Authority (APS) (Mooring station in Fig. 1). Additionally, due to the lack of in situ wave
direction (Dir) data, complementary data were obtained from the Global Ocean Waves Reanalysis
product through the Copernicus Marine Data Store platform (Copernicus Marine Service, n.d.), for the
period between April 14 and May 3, 2025, with a temporal resolution of 3 hours and a spatial resolution
of 0.083°, extracted from a point near the coast (location shown in Fig. 1).
Wind data for the period between April 14 and May 3, 2025, used in this study were obtained from the
Sistema de Informações da Qualidade do Ar (Qualar - Air Quality Information System; location in Fig. 1),
maintained by the Companhia Ambiental do Estado de São Paulo (CETESB - São Paulo State
Environmental Company).
Current direction and intensity were also provided by the APS (Mooring station in Fig. 1), with values for
four water column depths (4, 6, 8, and 10 m). Precipitation data for the analysis period were obtained
from the Paranapuã rain gauge station (location in Fig. 1), in the municipality of São Vicente, from the
Centro Nacional de Monitoramento e Alertas de Desastres Naturais (Cemaden - National Center for
Monitoring and Early Warning of Natural Disasters), with data collected between April 14 and May 3,
2025. Cumulative rainfall data, in millimeters, are recorded every 10 minutes during rainfall and every 60
minutes in the absence of precipitation. From these data, daily cumulative precipitation values were
calculated.
Tidal variation was estimated using the UTide package for Python (Codiga 2011). In situ sea level data
collected between August 2024 and November 2025 at Ilha das Palmas (24°00'31.0"S, 46°19'28.6"W,
Mooring station in Fig. 1) were used to determine the harmonic constants through UTide, which were
subsequently applied to generate predictions and interpret tidal phases for each of the analyzed dates
(Table 1).
Table 1
Acquisition times (UTC) of the images used and tidal information
Date Sensor Image acquisition Tide
Apr 15, 2025 Sentinel-2 13:12 Ebb, Neap
Apr 27, 2025 PlanetScope 13:37 and 13:38 Flood, Neap
Apr 30, 2025 PlanetScope 13:37 Ebb, Neap
May 2, 2025 Sentinel-2 13:12 Ebb, Neap
May 2, 2025 PlanetScope 13:44 Ebb, Neap
Page 5/24

2.3 Satellite image acquisition and processing
Satellite image selection and acquisition were based on the Sentinel-2 and PlanetScope sensors. For the
Sentinel-2 images, acquisition was carried out using the Copernicus Browser, a tool developed by the
European Union's space program that allows visualization and access to data from all Copernicus
missions. Sentinel-2 Level-1C Collection 1 was selected, which provides orthorectified top-of-
atmosphere (TOA) reflectance. The Sentinel-2 satellite carries a multispectral sensor (MultiSpectral
Instrument; MSI), with 13 spectral bands (with spatial resolution of 10, 20, and 60 meters), a 5-day revisit
time, and systematic global coverage over land masses and coastal areas up to 20 km from the coast
(European Space Agency, n.d). Two dates were selected: April 15, 2025 (pre-event) and May 2, 2025
(post-event).
PlanetScope TOA reflectance images were obtained from the Web imagery platform of the Programa
Brasil M.A.I.S (Accession Agreement No. 23/2025/SE/MJSP). This sensor stands out for its high
temporal (daily revisit) and spatial (3 m) resolution compared to other satellites. Three analysis dates
were selected: one pre-event (April 27, 2025) and two post-event (April 30 and May 2, 2025), to track the
extent of the sediment plume resulting from the mass movement. An additional PlanetScope image
from May 4 was used only to verify whether the plume was still detectable (not shown here).
To improve data quality and eliminate atmospheric effects, atmospheric correction of the images was
performed using the ACOLITE tool (version 20250402.0). The default configuration was used, based on
the "Dark Spectrum Fitting" approach, which performs particularly well in turbid, productive waters
(ACOLITE 2025). Cloud pixels were masked by ACOLITE using its default settings.
2.4 Estimation of turbidity, SMC, chlorophyll-a, and
transparency
Retrieval of the bio-optical descriptors of water quality was performed using the ACOLITE processor,
which, following atmospheric correction, generates Level 2R (L2R) surface reflectance products and,
from these, Level 2W (L2W) products corresponding to reflectance-derived parameters. The retrieval
algorithms were applied following the guidelines of the processor manual, developed for applications in
coastal and inland aquatic environments, including optically complex waters (Vanhellemont, 2023).
Chlorophyll-a concentration (chl, in mg/m³) was calculated using an algorithm based on the reflectance
ratio between the blue and green bands (Table 2), using calibration values from Franz et al. (2015) built
into the software. This methodology was adopted based on a study previously conducted in the area by
Fernandes et al. (2026, in press). The manual notes that this product should be interpreted with caution
in coastal waters due to interference from sediments and colored dissolved organic matter (CDOM).
Suspended matter concentration (SMC, in g/m³) was estimated using the algorithm of Nechad et al.
(2010; Table 2), using the red band (ACOLITE default). Turbidity (TUR, in FNU) was calculated using the
algorithm of Nechad et al. (2009). Secchi disk transparency depth (Zsd, in meters) was derived using the
Page 6/24

QAA-RGB method of Pitarch & Vanhellemont (2021). This parameter allows quantification of the
reduction in water clarity caused by sediment input following the event.
To characterize the spectral response of the water, remote sensing reflectance (Rrs) spectra were
extracted from the ACOLITE products at three points, one in each study region (waters adjacent to
Porchat Island, Santos Bay, and São Vicente Bay – Fig. 1), for all image dates. The points were selected
arbitrarily within each region, without a formal sampling design, and were used only for a qualitative
comparison of spectral shape and magnitude among dates, sensors, and regions.
Statistical processing of the spatial data was performed in QGIS (version 3.42.3), from which mean
values were extracted for the waters adjacent to Porchat Island and for Santos and São Vicente bays.
The extent of the sediment plume was delineated from the PlanetScope imagery, considering the SMC
products, where pixels above 15 g/m³ were selected to generate a binary raster (threshold chosen from
visual inspection of the plume boundary in the true-color images), which was then converted to a
polygon (shapefile) delimiting the area covered by the plume, allowing its area and dispersion dynamics
to be monitored. No in situ measurements of turbidity, SMC, chlorophyll-a, or Secchi depth were available
for this event, so the satellite-derived products could not be validated locally. Accordingly, the results are
interpreted as relative changes between dates and among regions, rather than as absolute
concentrations.
Page 7/24

Table 2
Parameters calculated from the satellite images.
Parameter Equation Notes
Chlorophyll-a R: logarithm of the ratio between the
maximum blue band reflectance and the
green band reflectance; a0: 0.2412; a1:
−2.0546; a2: 1.1776; a3: −0.5538; a4:
−0.4570.
Suspended
matter
concentration
: first calibration parameter; :
second calibration parameter; : third
calibration parameter ; W: red or near-
infrared band..
Turbidity AT: first calibration parameter; BT:
second calibration parameter; C: third
calibration parameter; W: red or near-
infrared band.
Secchi disk
depth Rrs: remote sensing reflectance;
band at which the spectral minimum of
Kd is located; Kd: vertical attenuation
coefficient.
Note: Based on , in order to eliminate potentially remaining biases, the polynomial S(ZSD,b) was
used to obtain the unbiased estimate.
3 Results
3.1 Water quality changes and plume extent
In the Sentinel-2 images, SMC, turbidity, and Zsd indicated a stronger response in the waters adjacent to
Porchat Island than in the bays between the pre-event (April 15) and post-event (May 2) dates (Table 3,
Fig. 2). Mean SMC and turbidity in these waters nearly doubled (from 11.15 to 21.09 g/m³ and from
11.40 to 21.14 FNU, respectively), and mean Zsd decreased from 2.01 to 1.12 m, whereas changes in
SVB and SB were smaller, indicating that the strongest response was concentrated near the island.
Chlorophyll-a increased in all areas, but not more strongly near the landslide site.
chl = 10(a0+a1+R+a2R2+a3R3+a4R4)
SMC = +BρAρρw1−ρw/Cρ Aρ Bρ
Cρ
ρ
TUR = +BTATρw1−ρw/C
ρ
ZSD =
log( )|0.14−Rrs(λm)|0.013
2.5Kd(λm)
λm :
ZSD
Page 8/24

Table 3
Mean values of water quality variables in the adjacent waters (AW) to Porchat Island, São Vicente Bay
(SVB) and Santos Bay (SB), derived from Sentinel-2 (S2, April 15 and May 2) and PlanetScope (April 27,
April 30 and May 2) images.
Variable Region S2 Apr
15
Planet Apr
27
Planet Apr
30
S2 May
2
Planet May
2
SMC (g/m³) AW 11.15 15.42 26.79 21.09 21.07
SVB 9.56 11.63 14.13 12.43 20.42
SB 8.51 10.59 14.67 10.02 11.71
Turbidity (FNU) AW 11.4 12.26 21.3 21.14 16.75
SVB 9.69 9.24 11.24 12.62 16.24
SB 8.63 8.42 11.66 10.12 9.31
Zsd (m) AW 2.01 – – 1.12 –
SVB 1.94 – – 1.3 –
SB 2.51 – – 2.11 –
Chlorophyll-a
(mg/m³)
AW 4.14 2 1.76 4.54 3.75
SVB 3.45 1.86 1.68 4.77 4.09
SB 3.68 1.93 1.72 4.33 3.88
In the PlanetScope images (Table 3, Fig. 3), SMC and turbidity increased in all regions on April 30, the
day after the event, and the increase was largest in the adjacent waters (SMC from 15.42 to 26.79 g/m³;
turbidity from 12.26 to 21.30 FNU). By May 2, the values in these waters had decreased but remained
above pre-event conditions, and those in SB had returned close to them (SMC of 11.71 g/m³ and turbidity
of 9.31 FNU, compared with 10.59 g/m³ and 8.42 FNU on April 27). In SVB, in contrast, SMC and turbidity
kept increasing until May 2 (to 20.42 g/m³ and 16.24 FNU), even though the plume mapped on April 30
and May 2 did not enter this bay. Chlorophyll-a decreased slightly in all regions on April 30 and increased
on May 2 to values higher than those of April 27 (Table 3). The values of April 30 should be interpreted
with caution, given the cloud cover in that image.
On May 2, the products of the two sensors showed the same spatial pattern and values of the same
order of magnitude. Mean SMC in the adjacent waters was practically identical (21.09 g/m³ in Sentinel-2
and 21.07 g/m³ in PlanetScope), and the largest difference occurred in SVB, where SMC was higher in
PlanetScope (20.42 g/m³) than in Sentinel-2 (12.43 g/m³). Differences in turbidity were at most 4.4 FNU,
and Sentinel-2 chlorophyll-a means were slightly higher than those of PlanetScope in all regions (by 0.45
to 0.79 mg/m³).
Page 9/24

The analysis of the PlanetScope images revealed rapid dispersion of the sediment plume generated after
the landslide (Fig. 4). It initially covered a total area of 7.5 km² on April 30 and 2.02 km² on May 2. By
May 4 (image not shown here, only discussed), the plume could no longer be identified in the previously
affected region. The plume was concentrated near Porchat Island and advanced over the central portion
of SB, without entering SVB.
3.2 Spectral response
The Rrs spectra obtained at the three extraction points showed, on all analyzed dates (Fig. 5), the
spectral shape typical of turbid coastal waters, with a reflectance peak between 530 and 580 nm and a
decline toward the near-infrared. On the dates preceding the pipeline rupture, the three sites showed
relatively similar Rrs magnitudes. On April 15, peaks ranged from 0.0105 sr⁻¹ (SVB) to 0.0157 sr⁻¹ (SB),
with the adjacent waters in an intermediate position (0.0127 sr⁻¹); on April 27, the adjacent waters and
SVB showed similar peaks (0.0163 and 0.0158 sr⁻¹), while SB recorded the lowest value (0.0108 sr⁻¹). On
April 30, a marked increase in Rrs was observed at all points, most notably in SB (0.0275 sr⁻¹), followed
by the adjacent waters (0.0255 sr⁻¹) and SVB (0.0185 sr⁻¹); on this date, however, the PlanetScope image
showed pronounced cloud cover, which should be taken into account in the analysis. On May 2, the
adjacent waters showed the highest Rrs peak (0.0215 sr⁻¹ in Sentinel-2 and 0.0183 sr⁻¹ in PlanetScope),
exceeding SB (0.0157 and 0.0125 sr⁻¹) and SVB (0.0092 and 0.0075 sr⁻¹). Good agreement was
observed between Sentinel-2 and PlanetScope regarding curve shape.
3.3 Meteoceanographic data
Regarding the meteoceanographic data, during the analyzed period, significant wave height (Fig. 6a)
ranged from 0.86 m to 2.59 m for the reanalysis data, and from 0.35 m to 2.5 m for the in situ data (R²
between the datasets of 0.73). Wave periods (Fig. 6b) ranged from 4.3 to 10.39 seconds (reanalysis
data) and from 4.1 to 24 seconds (in situ, with an R² of 0.57 between the datasets). Regarding
propagation direction (Fig. 6c), waves predominantly came from the Southeast (SE) quadrant until April
20, followed by a shift to South (S) until April 23, returning to the SE direction between April 24 and 28,
and remaining from the S thereafter.
Wind records obtained at the CETESB monitoring station (Fig. 6d), for the period from April 14 to May 3,
2025, indicated a predominance of weak to calm winds, with speeds ranging from 0 to 1.7 m/s. The
highest wind speeds (up to 1.7 m/s) occurred mainly between April 17 and 19, with additional peaks
recorded on April 30. Wind direction showed temporal variability, but with a predominance of winds from
the southeast quadrant (SE, between approximately 120° and 170°) during the afternoons, and from the
southwest/west-southwest quadrant (SW/WSW, between approximately 220° and 256°).
Regarding currents (Fig. 6e and f), the data showed a fairly similar pattern across the analyzed depths (4,
6, 8, and 10 m). Current speeds reached a maximum of 0.4 m/s. In terms of direction, a typically
Page 10/24

bidirectional pattern was observed, with currents predominantly oriented between the N-NE and S-SW
quadrants, with similar directions throughout the water column.
Precipitation data recorded at the Paranapuã station (Fig. 6g), in São Vicente, indicate the occurrence of
significant rainfall events during April 2025. On April 19, the highest volume was recorded, with 74.4 mm.
On the dates preceding the pipeline rupture (April 24, 25, and 28) cumulative rainfall of 20.8 mm, 16.4
mm, and 24.6 mm was recorded, respectively. On the day of the rupture (April 29), only 0.4 mm was
recorded, and dry conditions predominated in the following days.
4 Discussion
4.1 Drivers of the event
The integrated analysis of remote sensing data and meteoceanographic conditions indicates that the
mass movement event that occurred on Porchat Island constituted a point-source sediment input event
capable of altering the quality of the adjacent coastal waters, although its effects were predominantly
transient. The results indicate that the observed dynamics appear to have resulted from the interaction
between prior soil saturation, the sudden water input resulting from the pipeline rupture, and the
hydrodynamic processes acting in SB.
Episodic sediment input events triggered by infrastructure failures or mass movements on urbanized
slopes, although still poorly documented in the coastal remote sensing literature, find relevant parallels
in other regions of the world. On the California coast, Warrick et al. (2025) used satellite imagery
combined with hydrodynamic modeling to characterize fine sediment plumes originating from beach
nourishment operations, demonstrating that wave height, wind speed and direction, and sediment
settling velocity exert strong control over the direction and extent of the turbid plume, a
meteoceanographic control pattern very similar to that observed in this study for the dispersion of the
Porchat Island plume. In Taiwan, Zhou et al. (2026) reconstructed, through multisensor remote sensing,
a cascade of hydrogeological hazards triggered by a typhoon-induced landslide that dammed a river and
subsequently breached, showing how the integration of different orbital data sources allows detailed
temporal documentation of slope destabilization processes and their subsequent hydrological impacts,
reinforcing the potential of high-resolution remote sensing for monitoring episodic events similar to the
one in São Vicente.
Although only 0.4 mm of precipitation was recorded on the date of the rupture (April 29, 2025), the
rainfall data show that the terrain was already under the influence of conditions favorable to instability. In
the days preceding the event, cumulative rainfall of 61.8 mm over 5 days was recorded. This
precipitation history suggests that the soil may have presented a high degree of saturation, reducing its
mechanical resistance and increasing susceptibility to mass movements, which is consistent with the
pattern of disorderly urbanization on shallow residual soils and insufficient drainage systems already
Page 11/24

described for Porchat Island by Marques (2015) and with the history of anthropogenic pressure on the
Santos and São Vicente Estuarine System reported by Sartoretto (2014).
4.2 Oceanographic control and plume response
The dispersion dynamics of the sediment plume observed after the event on Porchat Island are
consistent with the prevailing meteoceanographic conditions in SB, especially with the wave and wind
regime recorded between April 29 and May 2, 2025. The initial propagation of the plume toward the
central portion of SB was mainly associated with the change in wave incidence direction. On April 29,
there was a rotation from the Southeast (SE) to the South (S) quadrant, with directions varying between
160° and 175°, favoring the transport of sediments remobilized in the Porchat Island region to adjacent
areas of the bay. This mechanism, combined with the peak significant wave height recorded during this
period (Fig. 6a), may help explain the extent observed on April 30, 2025, when the plume occupied its
largest mapped area, 7.50 km².
In the following days, however, the area occupied by the plume decreased rapidly, reaching 2.02 km² on
May 2, 2025, and becoming undetectable in the images from May 4. This contraction is consistent with
the evolution of meteoceanographic conditions over the period: significant wave height and wind speed,
which reached higher values around April 29–30, showed a declining trend in the following days (Fig. 6a,
d), with low to moderate wind speeds (up to 1.7 m/s) progressively reducing the energy available to keep
sediment material resuspended and actively mixed in the water column. With this decrease in energy, the
material in suspension tends to be progressively diluted by the continued action of local currents, which
showed a typically bidirectional pattern between the N-NE and S-SW quadrants throughout the water
column (Fig. 6e, f), favoring the exchange and renewal of the water mass in the region adjacent to the
island even under neap tide conditions (Green and Coco 2013). This pattern of a pulse followed by rapid
decay is consistent with the behavior described for other coastal plumes of episodic origin, controlled
mainly by wave height and wind speed and direction (Warrick et al. 2025).
The confinement of the plume to SB, without sediments entering SVB, is consistent with the base
circulation already described for the SES. Computational hydrodynamic modeling studies indicate that
the trajectories of continental waters reaching Santos Bay follow distinct preferential patterns depending
on their origin, for example, waters from the Cubatão River tend to reach Santos Bay preferentially
through the mouth of the São Vicente Estuary, while those from the Quilombo River preferentially follow
the Port Channel (Roversi et al. 2016), which helps explain why local circulation tends to
compartmentalize the dispersion of material between the different water bodies of the system, even in
the face of a point and abrupt input such as that from Porchat Island. This behavior is also consistent
with the long-term characterization of Santos Bay as a receiving environment for significant
contributions of suspended material, nutrients, and chlorophyll-a from the Santos and São Vicente
estuarine channels, especially during periods of higher river discharge, which favor eutrophication
processes in the bay (Moser et al. 2005).
Page 12/24

Chlorophyll-a was higher on May 2 than on the previous PlanetScope dates and than on April 15 in
Sentinel-2, reaching similar values in both sensors (about 3.8–4.8 mg/m³). However, Sentinel-2 values on
April 15 were already close to those of May 2 (3.5–4.1 mg/m³), whereas PlanetScope values on April 27
and 30 were lower (about 1.7–2.0 mg/m³), so the data indicate variability among dates rather than a
response clearly attributable to the landslide. Nutrient remobilization, runoff after the rainfall of April 19–
28, and the biogeochemical response described for this estuarine system (Moser et al. 2005) are
possible explanations, and interference of sediments and CDOM on the blue-to-green algorithm cannot
be excluded; in situ data would be needed to separate them. Nevertheless, SMC and turbidity in SVB
increased between April 27 and May 2 (Table 3), which may reflect continental runoff or estuarine
discharge after the April rainfall.
4.3 Sensor comparison and uncertainties
The Sentinel-2 and PlanetScope products for May 2 showed a consistent spatial pattern and values of
the same order of magnitude (Table 3). Mean SMC in the adjacent waters was practically identical
(21.09 and 21.07 g/m³), and the largest difference occurred in SVB (12.43 and 20.42 g/m³). The
PlanetScope maps showed more pronounced spatial variability, an effect consistent with its higher
spatial resolution, which preserves features of heterogeneous plumes that coarser pixels smooth out
(Dorji and Fearns 2017). Differences in acquisition time (about 30 minutes apart), spectral bands, and
radiometric calibration may also contribute, and without in situ data it was not possible to determine
which product is closer to the actual concentrations.
Comparisons between these sensors in other settings show that performance depends on the variable
and the algorithm. Mansaray et al. (2021) compared PlanetScope, Landsat-8, and Sentinel-2 for
chlorophyll-a, turbidity, and Secchi depth in agricultural reservoirs in Oklahoma (USA), and Wasehun et al.
(2025) found that the performance of Sentinel-2 and PlanetScope for chlorophyll-a in a North Carolina
reservoir varied with the algorithm. Regarding radiometric accuracy, Di Francesco et al. (2025) note that
the PlanetScope constellation lacks onboard radiometric calibration, which hinders corrections
compared to Sentinel-2, especially in scenes affected by sun glint and greater uncertainty in atmospheric
correction.
The increase in suspended solids associated with the resuspension and displacement of sediment by
waves and currents tends to raise turbidity and reduce water column transparency (Green and Coco
2013), as observed in the decrease of Zsd, because more suspended particles increase light scattering
and absorption and limit light penetration. The consequent reduction in the light available to
phytoplankton may limit primary production (Cloern 1987), but no clear chlorophyll-a response to the
event was detected here.
The Rrs spectra reinforce, with some caveats, the spatial pattern of the plume mapped via PlanetScope.
On May 2, under more favorable atmospheric conditions, the highest Rrs peak occurred in the adjacent
waters, consistent with the still elevated SMC and turbidity recorded there. The April 30 image, however,
requires greater caution. Although the generalized increase in Rrs at all points is compatible with the
Page 13/24

sediment input immediately after the rupture, this image had pronounced cloud cover. ACOLITE masked
cloud pixels, but cloud shadows over water are not removed by the processor and may have been
retrieved as water, biasing the products locally, and radiance from nearby clouds may have contaminated
neighboring water pixels through adjacency effects, artificially raising the reflectance (Feng and Hu
2016). Atmospheric correction is also more uncertain in scenes with high cloud cover (Pahlevan et al.
2021; Warren et al. 2019). Thus, the peak observed in SB (whose extraction point was outside the plume
mapped that day) cannot be safely attributed solely to sediment advection driven by the wave rotation
from SE to S on April 29. The same caution applies to the SMC, turbidity, and chlorophyll-a means of April
30 and to the plume area of 7.50 km² on that date, since masked areas may have concealed part of the
plume and shadows may have affected its delineation. Even so, the increase near the island on April 30
is consistent with the pattern observed in both sensors on May 2, so the direction of the response is
more reliable than its magnitude. This limitation reinforces the need to prioritize images with lower cloud
cover, and to mask cloud shadows, in the rapid assessment of similar events.
In addition, the Sentinel-2 pre-event image (April 15) precedes the heavy rainfall of April 19 (74.4 mm)
and April 24–28, so part of the changes observed in the bays may reflect continental runoff. The
PlanetScope baseline (April 27) was acquired after these rains, and in SVB the PlanetScope values kept
increasing until May 2 even though the mapped plume did not enter this bay, which suggests a
contribution from other sources.
4.4 Implications for coastal management and monitoring
The Porchat Island event illustrates a type of disturbance that is rarely captured by conventional
monitoring. The plume was detectable for only a few days, expanding to 7.50 km² by April 30 and
shrinking to 2.02 km² by May 2. Its maximum mapped extent could be documented only because
PlanetScope provides daily revisits at 3 m spatial resolution. The Sentinel-2 images available for the
analysis spanned the period of the event but did not coincide with its peak. On its own, therefore, this
sensor would have documented the change in water quality but not the extent or the rapid dissipation of
the plume, also given its coarser spatial resolution. This supports the argument of Dorji and Fearns
(2017) that sensor characteristics determine which features of heterogeneous plumes are resolved, and
suggests that, for episodic events, temporal resolution is as decisive as spatial resolution.
The event also shows how infrastructure and slope conditions interact with coastal water quality. The
rupture occurred on an urbanized slope after several rainy days, in an area classified as a Permanent
Preservation Area and monitored by civil defense for geological risk (Marques 2015). In this context, the
water released by the pipeline probably acted as a trigger on soil that may already have been close to its
stability limit, which suggests that risk assessments based only on rainfall thresholds may not capture
failures in drainage and water supply infrastructure. Studies of cascading effects triggered by landslides,
such as that of Zhou et al. (2026) in Taiwan, likewise emphasize the importance of combining
multisensor observations with information on triggering conditions. In the case of Porchat Island,
satellite products made it possible to document the coastal consequences of this failure, an aspect that
is often outside the scope of slope monitoring programs.
Page 14/24

The temporary nature of the optical changes should be interpreted with caution. The rapid dissipation of
the plume is consistent with the energetic dispersion and renewal of the water mass discussed, as well
as with the behavior of episodic plumes described by Warrick et al. (2025). However, the fate of the
approximately 860,000 m³ of mobilized material was not assessed. Sediment that left the water column
may have settled on the seafloor of Santos Bay, where its effects would not be visible in optical imagery.
In addition, the biological response, particularly the variation in chlorophyll-a among dates chlorophyll-a,
cannot be clarified without in situ data. Since Santos Bay supports uses that depend on water quality, the
disappearance of the surface signal cannot be equated with the absence of impacts. Rather, the results
indicate that satellite imagery is suitable for delineating the affected area and its short-term persistence,
whereas assessing ecological and deposition effects requires complementary field sampling soon after
similar events. Finally, the workflow used here relied on open-access Sentinel-2 data, institutional access
to PlanetScope, and standard ACOLITE processing, which makes it potentially replicable by
environmental agencies.
5 Conclusion
The mass movement triggered by the pipeline rupture on Porchat Island represented a point-source
sediment input with marked, though apparently transient, impacts on the water quality of the adjacent
coastal waters. Between April 15 and May 2, mean SMC in the adjacent waters increased by about 89%
(11.15 to 21.09 g/m³) and mean Secchi depth decreased by about 44% (2.01 to 1.12 m) in Sentinel-2,
with smaller changes in Santos and São Vicente bays. The sediment plume mapped with PlanetScope
decreased from a maximum mapped extent of 7.50 km² on April 30 to 2.02 km² on May 2, and was no
longer identifiable on May 4.
Plume dispersion was consistent with the combination of prior rainfall, which probably left the slope
saturated, the rotation of wave incidence toward the south quadrant, which directed the material toward
the central portion of Santos Bay, and the subsequent decline in wave and wind energy, which coincided
with the dilution of the plume by local currents. Chlorophyll-a did not show a response clearly
attributable to the event, since values varied among dates and sediment and CDOM interference on the
algorithm cannot be excluded.
These conclusions should be considered in light of some limitations. No in situ data were available for
validation. Therefore, results are interpreted as relative changes rather than absolute concentrations. The
Sentinel-2 April 15 image preceded the heavy rainfall of late April, so changes in the bays may partly
reflect runoff. The April 30 image was affected by cloud cover and cloud shadows; the number of post-
event images was limited, and the analysis of meteoceanographic forcing was qualitative. Within these
limitations, the combined use of Sentinel-2 and PlanetScope, and particularly the daily revisit of the latter,
made it possible to document the extent and persistence of the plume, demonstrating the potential of
high-resolution remote sensing for the rapid assessment of urban infrastructure failures and for
supporting coastal monitoring and management. The workflow relied on open-access Sentinel-2 data,
institutional access to PlanetScope, and ACOLITE processing, which makes it potentially replicable by
Page 15/24

environmental agencies. Future work combining rapid field sampling, hydrodynamic modeling, and
sensor-specific validation would help quantify sediment fate and biological effects.
Declarations
Author Contribution
T.O.A: Conceptualization; Methodology; Validation; Formal Analysis; Investigation; Data Curation; Writing
– Original Draft.B.S.F: Methodology; Software; Validation; Formal Analysis; Investigation; Data Curation;
Visualization; Writing – Original Draft.B.C.C: Conceptualization; Methodology; Writing – Review & Editing;
Supervision; Project Administration.
Acknowledgement
The authors express their sincere thanks to Mário da Silva Bueno Júnior (former deputy secretary of
SEMAN/SV) for his availability, collaboration, and for providing the data on the mass movement that
supported the preparation of this article. His contribution was fundamental to the development of the
analyses and to the construction of this work.
Data Availability
Data will be made available on request.
References

1. ACOLITE (2025) ACOLITE v20250402.0. GitHub.
   https://github.com/acolite/acolite/releases/tag/20250402.0

2. Arabi B, Salama MS, Pitarch J, Verhoef W (2020) Integration of multi-sensor satellite images and in-
   situ measurements for long-term water quality monitoring in coastal areas. Remote Sens Environ
   239:111632. https://doi.org/10.1016/j.rse.2020.111632

3. Barbier EB, Hacker SD, Kennedy C, Koch EW, Stier AC, Silliman BR (2011) The value of estuarine and
   coastal ecosystem services. Ecol Monogr 81(2):169–193. https://doi.org/10.1890/10-1510.1

4. Barbosa CCF, Novo EMLM, Martins VS (2019) Introdução ao sensoriamento remoto de sistemas
   aquáticos: princípios e aplicações. Instituto Nacional de Pesquisas Espaciais, São José dos
   Campos. http://www.dpi.inpe.br/labisa/livro/res/conteudo.pdf

5. Bende-Michl U et al (2011) Monitoring strategies and scale-appropriate hydrologic and
   biogeochemical modelling for natural resource management: conclusions and recommendations
   from a session held at the iEMSs 2008. Environ Model Softw 26(4):538–542.
   https://doi.org/10.1016/j.envsoft.2010.09.006
   Page 16/24

6. Boyce DG, Lewis MR, Worm B (2010) Global phytoplankton decline over the past century. Nature
   466(7306):591–596. https://doi.org/10.1038/nature09268

7. Brivio PA, Giardino C, Zilioli E (2001) Validation of satellite data for quality assurance in lake
   monitoring applications. Sci Total Environ 268(1):3–18. https://doi.org/10.1016/S0048-
   9697(00)00693-8

8. Cloern JE (1987) Turbidity as a control on phytoplankton biomass and productivity in estuaries. Cont
   Shelf Res 7(11–12):1367–1381. https://doi.org/10.1016/0278-4343(87)90042-2

9. Codiga DL (2011) The University of Rhode Island tidal analysis package. Graduate School of
   Oceanography, University of Rhode Island, Narragansett.
   https://doi.org/10.13140/RG.2.1.3761.2008

10. Coleman DJ et al (2022) Reconciling models and measurements of marsh vulnerability to sea level
    rise. Limnol Oceanogr Lett 7(2):140–149. https://doi.org/10.1002/lol2.10230

11. Copernicus Marine Service (n.d.) Copernicus Marine Data Store.
    https://data.marine.copernicus.eu/products?disc=facets

12. Dadson SJ et al (2003) Links between erosion, runoff variability and seismicity in the Taiwan orogen.
    Nature 426(6967):648–651. https://doi.org/10.1038/nature02150

13. Di Francesco S, Biondi F, Casentini B, Fazi S, Amalfitano S, D'Eugenio M, Todisco F, Casadei S,
    Giannone F (2025) Sentinel-2 and Planet-Scope as reliable tools for water quality monitoring of
    small reservoirs. Egypt J Remote Sens Space Sci 28(4):713–723.
    https://doi.org/10.1016/j.ejrs.2025.11.002

14. Dorji P, Fearns P (2017) Impact of the spatial resolution of satellite remote sensing sensors in the
    quantification of total suspended sediment concentration: a case study in turbid waters of Northern
    Western Australia. PLoS ONE 12(4):e0175042. https://doi.org/10.1371/journal.pone.0175042

15. European Space Agency (n.d) Sentinel-2.
    https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Sentinel-2

16. Fagherazzi S et al (2013) Marsh collapse does not require sea level rise. Oceanography 26(3):70–

17. https://doi.org/10.5670/oceanog.2013.47

18. Feng L, Hu C (2016) Cloud adjacency effects on top-of-atmosphere radiance and ocean color data
    products: a statistical assessment. Remote Sens Environ 174:301–313.
    https://doi.org/10.1016/j.rse.2015.12.020

19. Fernandes BS et al (2026) in press) Effects of different atmospheric corrections on the retrieval of
    water quality using ocean color remote sensing in an urbanized tropical estuarine system.
    Geoscience and Remote Sensing Letters

20. Fong CR, Gaynus CJ, Carpenter RC (2020) Extreme rainfall events pulse substantial nutrients and
    sediments from terrestrial to nearshore coastal communities: a case study from French Polynesia.
    Sci Rep 10(1):2955. https://doi.org/10.1038/s41598-020-59807-5

21. Franz BA, Bailey SW, Kuring N, Werdell PJ (2015) Ocean color measurements with the Operational
    Land Imager on Landsat-8: implementation and evaluation in SeaDAS. J Appl Remote Sens
    Page 17/24

9(1):096070. https://doi.org/10.1117/1.JRS.9.096070
21. G1 Santos (2025) Adutora da Sabesp rompe e causa deslizamento de terra na Ilha Porchat, em São
Vicente, SP. https://g1.globo.com/sp/santos-regiao/noticia/2025/04/29/adutora-da-sabesp-rompe-
e-causa-deslizamento-de-terra-na-ilha-porchat-em-sao-vicente-sp-video.ghtml
22. Gitelson AA, Schalles JF, Hladik CM (2007) Remote chlorophyll-a retrieval in turbid, productive
estuaries: Chesapeake Bay case study. Remote Sens Environ 109(4):464–472.
https://doi.org/10.1016/j.rse.2007.01.016
23. Gohin F et al (2019) Twenty years of satellite and in situ observations of surface chlorophyll-a from
the northern Bay of Biscay to the eastern English Channel: is the water quality improving? Remote
Sens Environ 233:111343. https://doi.org/10.1016/j.rse.2019.111343
24. Green MO, Coco G (2013) Review of wave-driven sediment resuspension and transport in estuaries.
Rev Geophys 52(1):77–117. https://doi.org/10.1002/2013RG000437
25. Hopkinson CS et al (2019) Chap. 1 - Coastal wetlands: a synthesis. In: Perillo GME et al (eds)
Coastal wetlands, 2nd edn. Elsevier, pp 1–75. https://doi.org/10.1016/B978-0-444-63893-9.00001-0
26. Irie M, Nakagawa A, Higashi T (2023) Quantitative evaluation of suspended solid runoff from large-
scale landslide areas presumed to be the source of turbid water. Water 15(18):3186.
https://doi.org/10.3390/w15183186
27. Jensen JR (2014) Remote sensing of the environment: an Earth resource perspective, 2nd edn.
Pearson, Upper Saddle River
28. Kirwan ML, Megonigal JP (2013) Tidal wetland stability in the face of human impacts and sea-level
rise. Nature 504(7478):53–60. https://doi.org/10.1038/nature12856
29. Kobiyama M et al (2011) Landslide influence on turbidity and total solids in Cubatão do Norte River,
Santa Catarina, Brazil. Nat Hazards 59(2):1077–1086. https://doi.org/10.1007/s11069-011-9818-4
30. Ladd CJT et al (2019) Sediment supply explains long-term and large-scale patterns in salt marsh
lateral expansion and erosion. Geophys Res Lett 46(20):11178–11187.
https://doi.org/10.1029/2019GL083315
31. Li P et al (2022) Spatiotemporal dynamics of suspended particulate matter in the Bohai Sea, China
over the past decade from the space perspective. Sci Total Environ 851(Pt 1):158210.
https://doi.org/10.1016/j.scitotenv.2022.158210
32. Lin GW et al (2011) Impact of rainstorm-triggered landslides on high turbidity in a mountain
reservoir. Eng Geol 117(1):97–103. https://doi.org/10.1016/j.enggeo.2010.10.009
33. Mansaray AS, Dzialowski AR, Martin ME, Wagner KL, Gholizadeh H, Stoodley SH (2021) Comparing
PlanetScope to Landsat-8 and Sentinel-2 for sensing water quality in reservoirs in agricultural
watersheds. Remote Sens 13(9):1847. https://doi.org/10.3390/rs13091847
34. Marques ACC (2015) Análise e considerações sobre o deslocamento (ruptura) de uma estrutura de
arrimo convencional em uma encosta da Ilha Porchat, São Vicente. Undergraduate thesis,
Universidade Estadual Paulista Júlio de Mesquita Filho, Guaratinguetá.
https://repositorio.unesp.br/entities/publication/3af309fa-55f9-4474-9c8a-bcb0a5f5cb70
Page 18/24

35. Mobley CD (1994) Light and water: radiative transfer in natural waters. Academic, San Diego

36. Moser GAO, Gianesella SMF, Barrera-Alba JJ, Bérgamo AL, Saldanha-Corrêa FMP, Miranda LB, Harari
    J (2005) Instantaneous transport of salt, nutrients, suspended matter and chlorophyll-a in the
    tropical estuarine system of Santos. Braz J Oceanogr 53(3–4):115–127.
    http://www.scielo.br/j/bjoce/a/rSyPJGwfMVhyJwGW5zfHVQJ/?lang=en

37. Nechad B, Ruddick K, Neukermans G (2009) Calibration and validation of a generic multisensor
    algorithm for mapping of turbidity in coastal waters. Proc SPIE 7473:74730H.
    https://doi.org/10.1117/12.830700

38. Nechad B, Ruddick K, Park Y (2010) Calibration and validation of a generic multisensor algorithm for
    mapping of total suspended matter in turbid waters. Remote Sens Environ 114(4):854–866.
    https://doi.org/10.1016/j.rse.2009.11.022

39. Neil DT et al (2002) Sediment yield and impacts from river catchments to the Great Barrier Reef
    lagoon: a review. Mar Freshw Res 53(4):733–752. https://doi.org/10.1071/MF00151

40. Pahlevan N et al (2021) ACIX-Aqua: a global assessment of atmospheric correction methods for
    Landsat-8 and Sentinel-2 over lakes, rivers, and coastal waters. Remote Sens Environ 258:112366.
    https://doi.org/10.1016/j.rse.2021.112366

41. Pitarch J, Vanhellemont Q (2021) The QAA-RGB: a universal three-band absorption and
    backscattering retrieval algorithm for high resolution satellite sensors. Development and
    implementation in ACOLITE. Remote Sens Environ 265:112667.
    https://doi.org/10.1016/j.rse.2021.112667

42. Rabalais NN et al (2007) Sediments tell the history of eutrophication and hypoxia in the northern
    Gulf of Mexico. Ecol Appl 17(sp5):S129–S143. https://doi.org/10.1890/06-0644.1

43. Roversi F, Rosman PCC, Harari J (2016) Análise das trajetórias das águas continentais afluentes ao
    Sistema Estuarino de Santos. Rev Bras Recur Hídricos 21(1):242–250.
    https://doi.org/10.21168/RBRH.V21N1.P242-250

44. Santa Portal (2025) Rompimento de adutora causa deslizamento de terra na Ilha Porchat.
    https://www.santaportal.com.br/ultimas-noticias/rompimento-de-adutora-causa-deslizamento-de-
    terra-na-ilha-porchat-video

45. Sartoretto JR (2014) Histórico de atividade antrópica no Sistema Estuarino Santos e São Vicente.
    Master's thesis, Universidade de São Paulo, São Paulo. https://repositorio.usp.br/item/002674381

46. Schaeffer BA et al (2013) Barriers to adopting satellite remote sensing for water quality
    management. Int J Remote Sens 34(21):7534–7544.
    https://doi.org/10.1080/01431161.2013.823524

47. Seiler LMN (2019) Hidrodinâmica e comportamento de variáveis indicadoras da qualidade da água
    no complexo estuarino de Santos-São Vicente-Bertioga: uma abordagem numérica. Doctoral thesis,
    Instituto Oceanográfico, Universidade de São Paulo, São Paulo.
    https://doi.org/10.11606/T.21.2020.tde-31012020-115011
    Page 19/24

48. Simão R et al (2025) Unprecedented suspended solids load caused by record-breaking extremes in
    southern Brazil: the May 2024 event. Front Mar Sci 12. https://doi.org/10.3389/fmars.2025.1726787

49. Smith MA et al (2021) Stormwater runoff and tidal flooding transform dissolved organic matter
    composition and increase bioavailability in urban coastal ecosystems. J Geophys Res Biogeosci
    126(7). https://doi.org/10.1029/2020JG006146. e2020JG006146

50. Vanhellemont Q (2023) ACOLITE user manual, QV – October 23, 2023. RBINS, Brussels

51. Volk M, Liersch S, Schmidt G (2009) Towards the implementation of the European Water Framework
    Directive? Lessons learned from water quality simulations in an agricultural watershed. Land Use
    Policy 26(3):580–588. https://doi.org/10.1016/j.landusepol.2008.08.005

52. Wang X et al (2022) How turbidity mediates the combined effects of nutrient enrichment and
    herbivory on seagrass ecosystems. Front Mar Sci 9. https://doi.org/10.3389/fmars.2022.787041

53. Warren MA et al (2019) Assessment of atmospheric correction algorithms for the Sentinel-2A
    MultiSpectral Imager over coastal and inland waters. Remote Sens Environ 225:267–289.
    https://doi.org/10.1016/j.rse.2019.03.018

54. Warrick JA, Stevens AW, Tehranirad B (2025) Coastal fine-grained sediment plumes from beach
    nourishment near Santa Barbara, California. Coast Eng J 67(3):558–582.
    https://doi.org/10.1080/21664250.2025.2497705

55. Wasehun ET, Hashemi Beni L, Di Vittorio CA, Zarzar CM, Young KRL (2025) Comparative analysis of
    Sentinel-2 and PlanetScope imagery for chlorophyll-a prediction using machine learning models.
    Ecol Inf 85:102988. https://doi.org/10.1016/j.ecoinf.2024.102988

56. Windham-Myers L, Crooks S, Troxler TG (eds) (2018) A blue carbon primer: the state of coastal
    wetland carbon science, practice and policy. CRC, Boca Raton.
    https://doi.org/10.1201/9780429435362

57. Zhou S, Li W, Li Y et al (2026) Remote sensing reconstruction of a typhoon-induced landslide–river
    damming–outburst flood cascade in the Matia'an River, Taiwan. Landslides 23:2349–2362.
    https://doi.org/10.1007/s10346-026-02772-3
    Figures
    Page 20/24

Figure 1
Location map of the study area. The polygon indicates the study area, subdivided into Santos Bay, São
Vicente Bay, and the waters adjacent to Porchat Island. Planet image from May 2, 2025.
Figure 2
Chlorophyll-a, SMC, Secchi disk transparency, and turbidity concentrations for Sentinel-2 images.
Page 21/24

Figure 3
Chlorophyll-a, SMC, and turbidity concentrations for Planet images. Two Planet images were used for
April 27.
Figure 4
Extent and dispersion of the sediment plume based on analysis of PlanetScope images in the days
following the event.
Page 22/24

Figure 5
Remote sensing reflectance obtained at each sampling point for the three analysis regions.
Page 23/24

Figure 6
Meteoceanographic conditions: (a) significant wave height (Hs); (b) period (T); (c) direction (Dir), gray
dots indicate Copernicus reanalysis data, black dots indicate in situ APS data; (d) wind speed and
direction; (e) current speed - in situ APS data; (f) current direction - in situ APS data; (g) daily
precipitation; red dashed lines indicate satellite image acquisition dates; n.d.a.: no data available.
Page 24/24