1. Introduction
The total suspended matter (TSM) concentration in aquatic systems is a key indicator of water quality and an essential factor for water environmental assessment [1,2]. TSM in estuaries originates from fluvial discharge and sediment resuspension due to coastal upwellings. Mainly, TSM fluctuations stem from transport, sedimentation, and resuspension [3]. The TSM’s influence extends to water clarity, color, sedimentary deposition, and the sedimentation process in estuarine and coastal areas [4,5,6], as well as the geomorphological evolution of such areas [7,8]. Thus, monitoring TSM changes is pivotal for coastal engineering, environmental conservation, and the development of ports and waterways [9,10]. Satellite remote sensing advancements have empowered scholars to pursue remote retrieval of TSM in various aquatic environments utilizing different data sets [11,12,13,14]. Initially, in the 1970s, research on TSM in estuarine environments relied on various data and methods. Weiblatt and Yarger pioneered using Landsat 1 MSS data to model statistical approaches to TSM retrieval [15]. Later, Williamson et al. [16] discovered a linear relationship between measured suspended sediment concentrations and remote sensing data during estuarine sediment concentration retrieval. Gordon et al. [17] formulated the well-known Gordon relationship based on diffuse reflectance scattering models and proposed an atmospheric correction method for Class I water bodies. Munday et al. [18] observed that the nonlinear relationship between radiance and suspended solids concentration offers a better curve fit than linear relationships. Tassan [19] constructed a three-component sea color algorithm using simulations, whereas Moore [20] correlated the blue/green band ratio with seasonal variations in estuarine turbidity, aiding future TSM remote retrieval studies. The evolution of remote sensing technology has facilitated a shift in water quality monitoring from qualitative to quantitative methods. Ramaswamy et al. [21] married SeaWiFS imaging data with ground-truth TSM measurements from Martaban Bay, forging an empirical high-turbidity region model. Li et al. [22] developed a slope-based algorithm for atmospheric correction over Class II waters for TSM retrieval. Liu et al. [23] created an empirical inversion model for TSM in the Pearl River Estuary via in situ TSM data and hyperspectral imaging, endorsing a three-band algorithm for precision.
In recent years, the use of remote sensing imagery for suspended solids concentration retrieval has been significantly expanded and refined. Kazemzadeh et al. [24] employed MODIS sensor imagery alongside measured suspended sediment concentrations to construct an artificial neural network (ANN) model. This was used to retrieve the suspended sediment concentration within the Bahmnshir River basin in southwestern Iran from 2003 to 2011 and monitor the basin’s spatiotemporal changes. Montanher [25] discovered that increased top atmosphere reflectivity could enhance the accuracy of the suspended sediment’s remote sensing inversion. In contrast, Dorji et al. [26] employed varied suspended sediment inversion models, based on MODIS and Landsat sensors, to Class I and Class II water bodies in northern Australia. They found that some high-accuracy retrieval models, when applied to new water bodies, resulted in significant errors. They suggested it is necessary to construct new retrieval models based on the study area’s characteristics when examining unknown water types.
Over time, remote sensing retrieval has been used to evaluate suspended solids’ spatiotemporal changes in various seas, rivers, and estuaries. However, existing algorithms developed for remote sensing retrieval of suspended solids are often site-specific or applicable to relatively limited concentration ranges, limiting their large-scale application [27]. The current research trend is geared towards establishing a suspended solids concentration retrieval system with broader applicability [28].
Liu et al. [29] proposed a comprehensive model for the quantitative remote sensing retrieval of suspended sediments, drawing on the physical mechanism of light propagation and a water pollution model developed by Deng et al. [30]. When applied to Landsat 5 images, the model demonstrated that the Pearl River Estuary’s suspended sediment distribution aligns with its underwater topography, with higher TSM concentration in the estuary mouth and west bank than the east. Subsequently, Xi et al. [31] established two-band empirical model for estimating TSM in the Pearl River Estuary, using MERIS band reflectance and measured data, and tested the model with the TSM data measured.
Numerous prior studies have demonstrated that the TSM in the Pearl River Estuary exhibits a seasonal pattern. Remote sensing inversion of the TSM across different seasons necessitates not only the use of various bands but also the employment of distinct inversion algorithms [32,33,34]. Consequently, a single, unified inversion model is insufficient for monitoring TSM in the Pearl River Estuary. During the wet season, factors such as increased precipitation, runoff, and seasonal upwelling contribute to the highest TSM levels in the Pearl River Estuary. High-value areas are predominantly concentrated in coastal regions, exhibiting a decreasing trend from nearshore to offshore areas. In contrast, during the dry season, the Pearl River Estuary’s seasonal upwelling weakens due to reduced runoff and precipitation, resulting in high-value areas primarily located near the estuary [34]. Liu [35] et al. developed a remote sensing retrieval model for suspended sediment using GF-1 satellite remote sensing data from the winter of 2019 and measured data from the Pearl River Estuary. The model aimed to monitor the Hong Kong–Zhuhai–Macao Bridge’s impact on the spatial distribution of suspended sediment. The findings revealed a significant difference of up to 7 mg/L in suspended sediment concentrations on either side of the bridge. Comparing the previously developed TSM models for the Pearl River Estuary unveiled significant differences with intertransformability and utilization being unachievable despite high spatiotemporal congruence. In other words, due to the limitations of satellite temporal and spatial resolution, the matching degree of ground-truth data, and the gaps in understanding the physical mechanisms of water movement in the Pearl River Estuary, there is currently no unified inversion model applicable to the Pearl River Estuary region. Seeking precision, this study selected satellite data with a high spatiotemporal match for ground-truth data and redrafted a TSM concentration retrieval model for the Pearl River Estuary.
2. Materials and Methods
2.1. Research Area and Measured Data
The Pearl River Estuary, colloquially known as Lingding Bay (21°48′~22°27′ N, 113°03′~114°19′ E), is embedded on the northern side of the South China Sea. This bell-shaped estuary covers an area of approximately 2000 km2 [36], making it a prominent feature of the region. Encompassed by the largest and most multifarious water ecosystem in southern China, the estuary comes to life through the contributions of three primary tributaries: the Xijiang, Beijiang, and Dongjiang Rivers. Flowing downstream, these water bodies converge to form an interlacing network of channels, bestowing upon the estuary a complex deltaic topology symbolically referred to as “three rivers converge and eight rivers diverge”. Situated within the subtropical monsoon region, the Pearl River basin’s climate oscillates between the alternating influences of the southwest monsoon in summer and the northeast monsoon in winter. In summer, the monsoon fetches tropical air masses, intensifying precipitation, leading to approximately 80% of the annual rainfall occurring in the wet season (April–September). Conversely, in winter, the prevailing cold, dry northeast monsoon drastically reduces precipitation, marking the dry season (October–March of the year that follows).
The morphology of this estuary classifies it as an irregular semi-diurnal tide with a minor tidal range, positioning it within the weak tide estuary category [37]. As shown in Figure 1, sediment transport in the Pearl River Estuary is a convoluted process, shaped by runoff, coastal currents, tidal forces, and other factors leaving a mark on the estuary. Sediments, guided by the Coriolis effect and the Ekman transport, hold a predominant southward and southwestward trajectory in the Pearl River estuary and adjacent areas [38,39]. Coupled with monsoon influences, surface and subsurface waters in the estuary experience turbulent mixing, promoting vertical sediment transport and amplifying the surface water’s TSM [40]. Tidal forces apply further temporal and spatial variability to the estuary’s sediment distribution. With the ascendance of the tide, the sediment-rich waters at the shoal and entrances are impeded and undergo northward spread along the west bank. When the tide recedes, the concentrated sediment disperses towards the open sea, tracing a trajectory extending southeast from the estuary [32]. Moreover, the west bank of the Pearl River Estuary consistently exhibits a higher TSM than its eastern counterpart. This spatial variation in sediment, heavily influenced by seasonality as well as precipitation and tidal forces, highlights the Pearl River Estuary’s dynamic nature, with a pronounced diffusion trend of suspended solids towards the open sea. As such, the concentration of suspended solids in the waters of the Pearl River Estuary is subject to significant seasonal fluctuations [41].
The data utilized in this paper, drawn from measurements of the Pearl River Estuary, were collected via navigation routes supported by two separate projects, resulting in certain discrepancies in the station data. The survey and research initiated in 2013 extended from Huangpu Port to the marine area near Guishan Island, encompassing Lingdingyang Bay and Bay Mouth. Within this expanse, concentrations of suspended solids in the Pearl River Estuary were obtained for both January and July of that year. The scope of station distribution expanded in May and November of 2020, as depicted in Figure 2, featuring a broader and more rational design.
Water quality samples were meticulously gathered in line with the regulations set forth by the “Marine Testing Standards.” Seawater was collected at a distance of 0.5 m from the surface. Samples were afterwards sealed and stored prior to being transported to the laboratory. There, the analysis and measurement of water quality were undertaken. Using the laboratory weighing technique, the mass concentration of suspended solids in the water samples from all four voyages was determined. This involved filtering the collected water samples with a filter membrane in the laboratory. Following high-temperature calcination, the mass of inorganic particles, or suspended solids, was measured and divided by the volume of the sample to acquire the mass concentration of suspended solids. The concentration range and average concentration values of total suspended solids from 54 effective sampling sites measured during four voyages are specifically shown in Table 1.
From the information in Table 1, it can be roughly discerned that there is a difference in the concentration of total suspended solids between the wet and dry seasons at the Pearl River Estuary, with the suspended solids concentration during the wet season being higher than that during the dry season every year.
2.2. Satellite Data Acquisition and Processing
The eighth release of the Landsat program, dubbed Landsat 8, was launched into orbit successfully on 11 February 2013 and began data reception in April the same year. Stationed in a sun-synchronous, polar orbit at 705 km with an inclination angle of 98.2°, Landsat 8 revisits its orbit every 16 days. The mechanics onboard consist of two instruments, namely the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). The OLI offers 9 bands each with a different spatial resolution referenced in Table 2 and an imaging swath of 185 × 185 km.
For this study, Landsat 8 spectral data were procured from the official website of the United States Geological Survey (USGS) via the
Prior to retrieval, water extraction is conducted using two popular methodologies: the Normalized Difference Water Index (NDWI) and the Modified Normalized Difference Water Index (MNDWI) computed using the formulas:
NDWI = (Green − NIR)/(Green + NIR)(1)
MNDWI = (Green − SWIR)/(Green + SWIR)(2)
where Green implies spectral reflectance in the green band, NIR that in the near-infrared band, and SWIR that in the shortwave infrared band.A comparative application of both methods on an identical image revealed their efficacy in extracting water body data from the Pearl River. Examining Figure 3a,b, we note the MNDWI method’s propensity to capture multiple minor water bodies, shifting the study’s focus from the Pearl River by blurring its boundary. Conversely, the NDWI method, limited by its inability to capture smaller water bodies, proves a suitable extraction methodology for this study. Weighing the strengths and weaknesses of both techniques, the NDWI method was chosen as the preferred water extraction algorithm for this research.
2.3. Remote Sensing Retrieval Algorithm
Turbidity retrieval algorithms commonly fall into three categories [42,43,44]: empirical, semi-analytical, and analytical algorithms. This study utilizes a semi-analytical algorithm, often referred to as a semi-empirical or semi-empirical, semi-analytical algorithm. Since 1990, this has been prevalent for monitoring water quality parameters. It has been specifically employed to retrieve the suspended sediment concentration in the Pearl River Estuary. It involves the analysis of spectral characteristics of remote sensing imagery in conjunction with the spectral properties of water quality parameters. A quantitative relationship between these two elements is established by optimally selecting bands or band combinations through mathematical methods [11,45]. A multitude of both domestic and international researchers have applied this algorithm for remote sensing retrieval of various water quality parameters (such as TSM, Chlorophyll a, yellow substances, transparency, and turbidity) in open waters, nearshore areas, lakes, and reservoirs, which yielded highly accurate retrievals [46,47,48].
The present study uses the remote sensing reflectance and the measured concentration of suspended solids in the sensitive bands with high correlation, as variables. This decision is based on the correlation between the visible reflectance data of Landsat 8 and the concentration of suspended solids in the Pearl River Estuary during both wet and dry seasons. Through regression analysis, a remote sensing retrieval model is established, reflecting the distribution of TSM in different periods and evaluating its spatiotemporal variations.
3. Results
3.1. Formation of a Remote Sensing Retrieval Model
Initially, an analysis is conducted for the correlation between the reflectance of Landsat 8 visible light bands (Bands 1–4) and the measured TSM in the Pearl River Estuary in both 2013 and 2020. Subsequently, suitable bands are identified for the remote sensing retrieval of the suspended matter concentration in the Pearl River Estuary for each period.
The interpretation of Figure 4 reveals a correlation coefficient of over 0.4 between the reflectance of each band and the TSM assessed in the wet season of 2013. Notably, the red and green bands (Band 3 and Band 4) exhibit sensitivity to the TSM concentration, with their correlation coefficients measuring 0.76 and 0.74, respectively. This implies heightened sensitivity to the TSM concentration in the Pearl River Estuary during the wet season; hence, these bands can be employed for subsequent modeling. In the dry season, the two blue bands (Band 1 and Band 2) show comparatively high correlation with the TSM concentration, where the correlation coefficients for Band 1 and Band 2 are 0.71 and 0.64, respectively. This marked sensitivity suggests their potential use as bands for retrieving the TSM concentration in the Pearl River Estuary during the dry season. Given the markedly low correlation of red and green bands with the TSM concentration, the dry season’s cleaner water may explain the diminished sensitivity to these bands and the heightened sensitivity to the blue band.
For the year 2020, the visible light band exhibits relative sensitivity to TSM concentration, where the correlation coefficient between each band and the TSM concentration was no less than 0.58. The red and green bands (Band 4 and Band 3) were distinguished as the most sensitive to TSM, with correlation coefficients of 0.88 and 0.87, respectively. They can hence be used as sensitive bands for formulating the TSM concentration retrieval model in the Pearl River Estuary during the wet season. Similarly, during the dry season, red and green bands depict sensitivity to the TSM, where their correlation coefficients are 0.77 and 0.72, respectively. They can be employed as sensitive bands for modeling and analyzing TSM distribution in the Pearl River Estuary during the dry season.
In order to model the TSM concentration during high- and low-water periods, this research established four different retrieval models: linear, quadratic, exponential, and power exponent, all of which utilized the selected sensitive bands. The selection of the optimal model was determined by using the associated determination coefficients of the retrieval model, as illustrated in Figure 5, Figure 6, Figure 7 and Figure 8.
3.2. Results and Error Analysis of TSM in the Pearl River Estuary
Based on the chosen secondary remote sensing retrieval model, Figure 9a illustrates the distribution of the suspended sediment concentration within the Pearl River Estuary. It is evident that the estuary’s proximity and mouth contain high concentrations of suspended sediment, with values ranging from 120 to 180 mg/L and low values between 15 and 20 mg/L. Furthermore, the concentration of suspended sediment appears to be higher near the coast and lower offshore, displaying a distribution pattern where the west bank has higher values compared to the east bank.
The remote sensing retrieval results of TSM during the dry season of the Pearl River Estuary, as demonstrated in Figure 9b, reveal high concentrations near the estuary and mouth, with values ranging from 50 to 180 mg/L and low values between 15 and 20 mg/L. It is evident that the overall suspended sediment concentration is lower than during the wet season. Regarding spatial distribution, the dry season’s concentration also exhibits higher values near the coast and lower values offshore. However, the distribution of clean water is more expansive compared to the wet season.
The wet season remote sensing retrieval results for TSM in the Pearl River Estuary in 2020, as shown in Figure 10a, range from 5 to 52 mg/L, which is lower compared to 2013. The concentration near the estuary and mouth is higher than in the offshore central area, with higher values on the west bank compared to the east bank. This forms a belt-like distribution stretching from northwest to southeast. The selected spectral remote sensing retrieval model presents the dry season remote sensing retrieval results in Figure 10b, exhibiting a range of 6–30 mg/L, which is lower than in the wet season. The spatial distribution is generally consistent with the wet season’s distribution characteristics, with higher concentrations near the estuary and mouth, and a belt-like pattern extending from northwest to southeast.
To verify the accuracy and precision of the remote sensing retrieval model for TSM in the Pearl River Estuary, an error analysis was conducted. Comparisons and the error analysis were performed between the remote sensing retrieval results of TSM and actual TSM measurements at various stations within the Pearl River Estuary, as depicted in Figure 11. The comparison illustrates a strong correlation between measured values and retrieval results, indicating the reliability of the retrieval outcomes.
4. Discussion
4.1. The Differences and Reasons for the TSM between the Flood Season and Dry Season
The remote sensing data for TSM during the flood and dry seasons in the Pearl River Estuary for the years 2013 and 2020 indicate that there is a strong seasonal pattern in the distribution of TSM. This pattern reveals higher concentrations in the flood season compared to the dry season, which can be attributed to seasonal variation in precipitation and runoff levels. Providing an illustration, the flood season, spanning April to September, accounts for around 70–85% of the annual precipitation and 78% of the annual runoff. Conversely, the dry season, from October to March of the succeeding year, sees runoff account for approximately 22% of the annual total, as per the Water Resources Department of the Pearl River Water Resources Commission.
During the flood season, an augment in river discharge from the Pearl River results in a greater influx of terrestrial materials. Moreover, an elevation in precipitation levels leads to a stronger freshwater input at the river mouth, instigating an upwelling current and resuspension of sediment particles [49]. The Ekman theory posits that the influence of the summer monsoon impels a seasonal upwelling on the southwest side of the river delta in the Pearl River Estuary [50]. This upwelling action resuspends bottom sediments to the water surface, subsequently raising the TSM concentrations in the estuary.
Contrastingly, the dry season sees a depletion in the Pearl River’s discharge levels, culminating in a reduction in the suspended sediment concentration and sediment transport capacity. This lessens the erosion of the estuary mouth’s bottom sediments, gradually diminishing the spatial distribution of suspended sediment concentration from the nearshore towards offshore areas. Further, the influence of the northeast monsoon means the freshwater input fails to accumulate sufficiently at the river mouth. This leads to the emergence of areas with a higher concentration near each estuarine outlet.
4.2. Differences and Reasons Analysis for TSM between 2013 and 2020
Figure 9 and Figure 10 depict a noticeable disparity in the TSM of the Pearl River Estuary between 2013 and 2020. In 2013, the highest TSM concentration reached 80 mg/L during the dry season, whereas the highest concentration in the flood season of 2020 was merely 52 mg/L. To ascertain the causes for these differences, this study chiefly concentrates on a three-pronged analysis, focusing on natural and anthropogenic factors.
4.2.1. Influence of Precipitation
The Pearl River Basin can be segregated into four major water resource zones: the East River, the West River, the North River, and the Pearl River Delta (as depicted in Figure 1). As each tributary’s precipitation in the Pearl River Basin affects the Pearl River Delta, it becomes necessary to separately calculate the precipitation for each river basin.
As per the 2013 Water Resources Bulletin promulgated by the Pearl River Water Resources Commission of the Chinese Ministry of Water Resources, the Pearl River region, encompassing the Pearl River Basin, international rivers situated east of—but not including—the Lancang River, the Han River, rivers in eastern Guangdong Province, coastal rivers in southwestern Guangxi Province, and rivers on Hainan Island as well as other South China Sea islands, registered a precipitation depth amounting to 1744.5 mm. This translates to a precipitation volume of roughly 10,080.7 billion m3. The bulletin underscores an increase of 5.1% compared to the previous year, 2012, and an increase of 12.7% more than the average annual precipitation. The report concludes that the year in question can be characterized as a year of substantial water availability. Conversely, the 2020 Water Resources Bulletin of Guangdong Province reports that the average annual precipitation in the province was 1574.1 mm, equating to a total precipitation volume of 2795.2 billion m³ for the year, which is 11.1% below average, suggestive of a relatively dry year.
Inspection of Figure 12 reveals that the total precipitation for 2013 noticeably exceeded both the average yearly precipitation and the total precipitation for 2020. Measured at the West River, North River, East River, and Pearl River Delta, precipitation levels in 2013 reached 2023, 2055.1, 2070.8, and 2176.7 billion m³, respectively, reflecting increases of 24.5%, 16.5%, 19.5%, and 18.0% compared to the corresponding average levels. Conversely, 2020 saw a precipitation decrease at the West River and the East River amounting to 11.1% and 16.5%, respectively, whilst the North River and the Pearl River Delta experienced a marginal increase of 4.2% and a decrease of 4.2%, respectively.
As seen in Figure 12, which compares annual precipitation in varying zones of the Pearl River Basin in 2013 and 2020, it is apparent that 2013 benefited from rich precipitation, signifying a year rich in water resources. On the other hand, 2020 garnered less precipitation than the average, denoting a relatively arid year. Total precipitation in the Pearl River Basin amounted to 8325.6 billion m³ in 2013 and just 6491.8 billion m³ in 2020—a reduction of 22.03% from 2013. This significantly influences the suspended sediment concentration in the Pearl River Estuary, reaching a high of 180 mg/L in the wet season and a peak at 80 mg/L in the dry season in 2013. Comparatively, in 2020, the suspended sediment concentration at the height of the wet season was just 52 mg/L and, in the dry season, only 30 mg/L.
4.2.2. Influence of Runoff Volume
In 2013, there was a notable surge compared to the longstanding average; this surge ranged between 18.6% and 25.3% across various river basins. The Western River marked the greatest rise at 25.3%, whilst the Northern River marked the smallest rise at 18.6%. The Eastern River and Pearl River Delta also marked increases of 20.5% and 18.8%, respectively. However, in 2020, a runoff volume decrease was reported at the Dongjiang River Basin and Pearl River Delta of 19.2% and 7.7%, respectively, despite slight increases reported at the Xijiang and Beijiang Rivers of 8.0% and 4.8%, respectively, against their long-term averages.
The historical data, as depicted in Figure 13, illustrate a positive correlation between precipitation and runoff volume. It is observed that an increase in rainfall broadly results in analogous augmentation of the runoff volume across distinct zones within the Pearl River Basin. In the year 2020, the Pearl River Basin recorded a total rainfall of 649.18 billion m3, representing a drop of 22.03% relative to the year 2013. This decrease correspondingly affected the total amount of runoff, registering 164.46 billion cubic meters in 2020, a 17.76% decrease from 199.98 billion cubic meters in 2013. These findings indicate the runoff volume’s influence on the TSM in the Pearl River Estuary is parallel and complementary to the precipitation’s effect. Both factors are interlinked and should not be viewed independently for any meaningful discussion.
By conducting a statistical analysis of the precipitation and runoff volume, coupled with a qualitative analysis of their correlation with the TSM within the Pearl River Estuary, it is discovered that elevated precipitation results in the accumulation of pronounced sediment from land sources within the waterways. This causes an increase in the runoff volume and intensifies the sediment content in the water. A rising runoff volume increases the hydraulic potential energy within the river, thereby enhancing its effect on the sediment at the river mouth and inducing sediment resuspension in the estuary. Concurrently, the river mouth discharges water laden with significant sediment, accounting for the higher concentration during the flood season as opposed to the dry season and its wider spatial distribution. The factors causing spatial and temporal variations in TSM within the Pearl River Estuary are, predominantly, of natural origin.
4.2.3. The Impact of Water Consumption
Assessing human factors in the Pearl River Estuary basin, such as water consumption and water types, reveals that, although total water consumption in the basin is noteworthy, it does not reach the scale of overall runoff and precipitation, as shown in Figure 14. Moreover, its impact on TSM is not comparable. Therefore, human activities exert limited influence on the overall suspended sediment distribution in the Pearl River Estuary. Nonetheless, our study identified that since 2013, the eastern part of the Pearl River Delta, specifically Shenzhen Bay, has consistently exhibited a high TSM during both high- and low-water level periods, particularly at the mouth of the Shenzhen River.
As demonstrated in Figure 15, the TSM in Shenzhen Bay during the 2013 flood season was recorded between 100 and 180 mg/L. During the dry season, the higher values ranged from 40 to 50 mg/L, primarily on the north side. In 2020, the higher concentration values of the suspended sediment during the flood season were 15–17 mg/L, while in the dry season, the values concentrated between a range of 21 and 24 mg/L. The high-concentration areas were located at the Shenzhen River mouth and the southern part of Shenzhen Bay. The suspended sediment concentration in the Shenzhen River mouth remained within 15–25 mg/L during these periods.
Subsequent investigations and review of pertinent information indicated that numerous factories situated near the Shenzhen River coincide with aquaculture locales and port terminals in Shenzhen Bay. These findings suggest a continual high concentration of suspended sediment, especially near the Shenzhen River mouth, potentially attributed to human industrial and agricultural interventions. Additionally, the scenario in Shenzhen Bay implies that the impacts of human activities tend to be localized, fleeting, and confined to certain areas in comparison to natural factors. Only extensive, prolonged human activities within a specific region can influence the spatial and temporal distribution of suspended sediment concentrations on a broader scale.
4.3. The Impacts of the Hong Kong–Zhuhai–Macao Bridge on the Pearl River TSM
The Hong Kong–Zhuhai–Macao Bridge, a significant bridge-tunnel project connecting Hong Kong, Zhuhai, and Macao, is situated in the Pearl River Estuary. Construction work commenced on 15 December 2009; 7 May 2013 marked the connection of the first underwater tunnel section to the artificial island, and the main structure of the bridge was finalized on 7 July 2017. The bridge passed the main project’s acceptance inspection on 6 February 2018.
The bridge, originating from an artificial island near the Hong Kong International Airport and connecting to the artificial islands of Zhuhai and Macao, terminates at the Hongwan Interchange in Zhuhai. It boasts a total length of 55 km, incorporating 22.9 km of bridge work and a 6.7 km underwater tunnel connecting the eastern and western artificial islands. The bridge comprises 224 piles, seven towers, and spans 33.1 m. The underwater tunnel stretches 5664 m, is 28.5 m wide, and 5.1 m high. With its longest underwater tunnel, the Hong Kong–Zhuhai–Macao Bridge currently stands as the longest bridge worldwide and a testament to modern China’s remarkable achievements.
To evaluate the impact of the Hong Kong–Zhuhai–Macao Bridge on the Pearl River Estuary’s TSM, the time of the highest observed concentration during the flood season was chosen. In 2013 (Figure 16), during the bridge’s construction phase, the bridge’s effect on the Pearl River Estuary’s TSM distribution was relatively minor. However, the bridge’s interception effect on the west bank of the Lingdingyang Estuary was observed, resulting in a higher concentration on the bridge’s northern side compared to the south. TSM levels on the northern side of the bridge in this area exceeded 100 mg/L, while on the southern side, they generally remained below 100 mg/L. Notably, in the eastern part of the bridge, between the two artificial islands in the underwater tunnel section, a localized increase in TSM was observed, potentially due to the resuspension of underwater sediment triggered by the tunnel’s construction.
The main construction of the Hong Kong–Zhuhai–Macao Bridge was completed in 2018, and it has been in operation since 2020. Over time, the impact of the bridge on the distribution of suspended solids in the Pearl River Estuary has become increasingly apparent. As shown in Figure 17, it is clear that the Hong Kong–Zhuhai–Macao Bridge significantly intercepts suspended solids. The northern part of the bridge exhibits a higher concentration of suspended solids than its southern counterpart, with sediment accumulation observed to spread from west to east. Even during the flood season of 2020, when the concentration of suspended solids was not high, the northern area near the bridge had a concentration that was 1–4 mg/L higher than the southern area. Additionally, the spatial distribution of suspended solids in the underwater tunnel section, located in the eastern part of the bridge, remains largely unaffected.
5. Conclusions
This study utilized Landsat 8 OLI multispectral data and measured remote sensing TSM data to establish optimal remote sensing retrieval models for suspended solids concentration during the flood and dry seasons in the Pearl River Estuary in 2013 and 2020. The obtained retrieval results helped derive the distribution of TSM and analyze the influence of natural and human factors on the distribution and spatiotemporal characteristics in the area. Our findings indicate that the suspended solids concentration in the Pearl River Estuary displays notable seasonality, with lower concentrations during the dry season and higher concentrations during the flood season. The spatial distribution of the suspended solids concentration is generally characterized by higher concentrations near the coast and lower concentrations offshore, decreasing from northwest to southeast due to runoff and tidal effects.
Considering the peak concentrations of TSM during different periods in 2013 and 2020, the influence of natural factors such as the total rainfall and runoff in the Pearl River Basin is very evident. Compared to the pervasive influence of natural factors, human activities have a localized and temporary effect. Only long-term and large-scale human activities can impact the distribution of the suspended solids concentration. The study also highlights the growing impact of the large-scale maritime construction project, the Hong Kong–Zhuhai–Macao Bridge, on the distribution of the suspended solids concentration over time.
Future research will focus on the quantitative analysis of the relationship between runoff, precipitation, and the distribution of the suspended solids concentration in the estuary. Additionally, ongoing monitoring of the Hong Kong–Zhuhai–Macao Bridge’s impact on the suspended solids concentration in the Pearl River Estuary will be carried out.
Conceptualization, Z.M. and Y.Z. (Yuanzhi Zhang); methodology, J.F.; software, J.Y.T. and Y.Z. (Yong Zhao); validation, W.Z., Y.Z. (Yong Zhao), and Y.L.; formal analysis, Z.M., J.F., and Y.Z. (Yong Zhao); investigation, W.Z. and Y.Z. (Yuanzhi Zhang); resources, J.Y.T.; data curation, W.Z. and Y.Z. (Yuanzhi Zhang); writing—original draft preparation, Z.M.; writing—review and editing, Y.Z. (Yuanzhi Zhang) and J.Y.T.; visualization, W.Z., J.F., and Y.L.; supervision, Y.Z. (Yuanzhi Zhang); project administration, Y.Z. (Yuanzhi Zhang); funding acquisition, Y.Z. (Yong Zhao). All authors have read and agreed to the published version of the manuscript.
Not applicable.
Not applicable.
Data are contained within the article.
The Landsat data from NASA and in situ measurements from local governments are highly appreciated.
The authors have no conflicts.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 1. Geographical location and distribution of tributaries in the Pearl River Estuary (Landsat 8 non-standard pseudo-color composite image).
Enlarge this image.Figure 2. Distribution map of actual shipborne survey sampling points in the Pearl River Estuary: (a) Sampling points in January 2013; (b) Sampling points in July 2013; (c) Sampling points in May 2020; (d) Sampling points in November 2020.
Enlarge this image.Figure 3. Comparison of extraction results between two water extraction methods: (a) Modified Normalized Difference Water Index; (b) Normalized Difference Water Index.
Enlarge this image.Figure 4. Correlation analysis between measured TSM and various visible light wavelengths.
Enlarge this image.Figure 5. The bimodal (Band 3 + Band 4) secondary remote sensing inversion model for the 2013 flood season.
Enlarge this image.Figure 6. The bimodal (Band 1/Band 2) secondary remote sensing inversion model for the 2013 dry season.
Enlarge this image.Figure 7. The bimodal (Band 3 + Band 4) secondary remote sensing inversion model for the 2020 flood season.
Enlarge this image.Figure 8. The unimodal (Band 3) secondary remote sensing inversion model for the 2020 dry season.
Enlarge this image.Figure 9. Spatial distribution map of TSM concentration in the Pearl River Estuary in 2013: (a) During the flood season, (b) During the dry season.
Enlarge this image.Figure 10. Spatial distribution map of TSM concentration in the Pearl River Estuary in 2020: (a) During the flood season, (b) During the dry season.
Enlarge this image.Figure 11. Comparison between TSM obtained from the remote sensing retrieval model and measured TSM at stations in the Pearl River Estuary.
Enlarge this image.Figure 12. Comparison of annual precipitation in different zones of the Pearl River Basin in 2013 and 2020. (The mean annual precipitation (including water resources) is based on the average values from 1956 to 2000, as stipulated in the “Technical Regulations for Water Resources Investigation and Evaluation.” The average annual precipitation is calculated using the weighted average method based on precipitation data from rain gauge stations in each zone.).
Enlarge this image.Figure 13. Comparison of annual runoff in different tributaries of the Pearl River Basin in 2013 and 2020 with the long-term average runoff.
Enlarge this image.Figure 14. Comparison of precipitation, runoff, and water usage in the Pearl River Basin in 2013 and 2020.
Enlarge this image.Figure 15. Distribution of TSM in Shenzhen Bay: (a) During the 2013 flood season; (b) During the 2013 dry season; (c) During the 2020 flood season; and (d) During the 2020 dry season.
Enlarge this image.Figure 16. Concentration distribution of TSM near the Hong Kong–Zhuhai–Macao Bridge during the 2013 flood season.
Enlarge this image.Figure 17. Concentration distribution of TSM near the Hong Kong–Zhuhai–Macao Bridge during the 2020 flood season.
Measured suspended matter concentration data from the voyages.
| Time | Total Station | Effective Station | TSM (mg/L) | Average (mg/L) |
|---|---|---|---|---|
| January 2013 | 15 | 13 | 9.80–18.40 | 13.39 |
| July 2013 | 15 | 12 | 5.49–64.19 | 47.94 |
| May 2020 | 19 | 14 | 4.40–32.40 | 13.24 |
| November 2020 | 20 | 15 | 3.20–24.10 | 9.14 |
Landsat 8 OLI sensor parameters.
| Band | Wavelength (μm) | Spatial Resolution (m) |
|---|---|---|
| Band 1 Coastal | 0.433–0.453 | 30 |
| Band 2 Blue | 0.450–0.515 | 30 |
| Band 3 Green | 0.525–0.600 | 30 |
| Band 4 Red | 0.630–0.680 | 30 |
| Band 5 NIR | 0.845–0.885 | 30 |
| Band 6 SWIR 1 | 1.560–1.660 | 30 |
| Band 7 SWIR 2 | 2.100–2.300 | 30 |
| Band 8 Pan | 0.500–0.680 | 15 |
| Band 9 Cirrus | 1.360–1.390 | 30 |
References
1. Webster, T.; Lemckert, C. Sediment resuspension within a microtidal estuary/embayment and the implication to channel management. J. Coast. Res.; 2002; 36, pp. 753-756. [DOI: https://dx.doi.org/10.2112/1551-5036-36.sp1.753]
2. Lv, J.W.; Liu, X.N.; Wang, J.; Cheng, J.B. Remote sensing retrieval of the suspended solid concentration in coast of South China based on PSO_RBF. Mar. Environ. Sci.; 2013; 32, pp. 669-673.
3. Chen, S.; Zhang, G.; Yang, S.; Yu, Z. Temporal and Spatial Changes of Suspended Sediment Concentration and Resuspension in the Yangtze River Estuary and Its Adjacent Waters. Acta Geogr. Sin.; 2004; 59, pp. 260-266.
4. Chen, J.Y.; Chen, S.L. Estuarine and coastal challenges in China. Mar. Geol. Lett.; 2002; 18, pp. 1-5+3.
5. Zhang, X.Y.; Pan, D.L. Study on Coastal Water Quality Evaluation Based on Ocean Color Remote Sensing Satellite. Proceedings of the 6th Symposium on Imaging Spectroscopy Technology and Applications; Shanghai, China, 28 October 2006.
6. Pan, D.L.; Gong, F. Progress in Application Technology of Satellite Ocean Remote Sensing in China. J. Hangzhou Norm. Univ.; 2011; 10, pp. 1-10.
7. Barnes, B.B.; Hu, C.; Kovach, C.; Silverstein, R.N. Sediment plumes induced by the Port of Miami dredging: Analysis and interpretation using Landsat and MODIS data. Remote Sens. Environ.; 2015; 170, pp. 328-339. [DOI: https://dx.doi.org/10.1016/j.rse.2015.09.023]
8. Cai, L.; Tang, D.; Levy, G.; Liu, D. Remote sensing of the impacts of construction in coastal waters on suspended particulate matter concentration-the case of the Yangtze River delta, China. Int. J. Remote Sens.; 2016; 37, pp. 2132-2147. [DOI: https://dx.doi.org/10.1080/01431161.2015.1121302]
9. Hunter, K.A.; Liss, P.S. The surface charge of suspended particles in estuarine and coastal waters. Nature; 1979; 282, pp. 823-825. [DOI: https://dx.doi.org/10.1038/282823a0]
10. Ma, H.; Zhong, H.L.; Luo, Z.; Zhu, L.; Shi, F. Suspended sediment concentrations in the Yangtze Estuary based on Landsat 8 remote sensing retrieval. Shanghai Land Resour.; 2018; 39, pp. 80-84.
11. Morel, A. In-water and remote measurements of ocean color. Bound.-Layer Meteorol.; 1980; 18, pp. 177-201. [DOI: https://dx.doi.org/10.1007/BF00121323]
12. Dekker, A.G.; Peters, S.W. The use of the Thematic appear for the analysis of eutrophic lakes: A case study in the Netherlands. Int. J. Remote Sens.; 1993; 14, pp. 799-822. [DOI: https://dx.doi.org/10.1080/01431169308904379]
13. Zhang, Y.Z.; Nie, Y.P.; Lin, Q.Z.; Jing, L.-H.; Zhang, B. Surface Water Quality Monitoring Using Remote Sensing. Remote Sens. Technol. Appl.; 2000; 4, pp. 214-219.
14. Kloiber, S.M.; Brezonik, P.L.; Olmanson, L.G.; Bauer, E.M. A procedure for regional lake water clarity assessment using Landsat multispectral data. Remote Sens. Environ.; 2002; 82, pp. 38-47. [DOI: https://dx.doi.org/10.1016/S0034-4257(02)00022-6]
15. Wang, Q. Spatial and temporal distribution of suspended sediment in offshore region of Haiyang city by remote sensing. Master’s degree; University of Chinese Academy of Sciences: Qingdao, China, June 2019.
16. Williamson, A.N.; Grabau, W.E. Sediment concentration mapping in tidal estuaries. NASA Spec. Publ.; 1974; 351, pp. 13-47.
17. Gordon, H.R. Remove of atmospheric effects from satellite imagery of the oceans. Appl. Opt.; 1978; 17, pp. 1631-1636. [DOI: https://dx.doi.org/10.1364/AO.17.001631] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20198035]
18. Munday, J.C.; Alföldi, T.T. Landsat test of diffuse reflectance models for aquatic suspended solids measurement. Remote Sens. Environ.; 1979; 8, pp. 169-183. [DOI: https://dx.doi.org/10.1016/0034-4257(79)90015-4]
19. Tassan, S. An improved in-water algorithm for the determination of chlorophyll and suspended sediment concentration from Thematic Mapper data in coastal waters. Int. J. Remote Sens.; 1993; 14, pp. 1221-1229. [DOI: https://dx.doi.org/10.1080/01431169308904406]
20. Moore, K.A.; Wetzel, R.L.; Orth, R.J. Seasonal pulses of turbidity and their relations to eelgrass (Zostera marina L.) survival in an estuary. J. Exp. Mar. Biol. Ecol.; 1997; 215, pp. 115-134. [DOI: https://dx.doi.org/10.1016/S0022-0981(96)02774-8]
21. Ramaswamy, V.; Rao, P.S.; Rao, K.H.; Raiker, V. Tidal Influence on Suspended Sediment Distribution and Dispersal in the Northern Andaman Sea and Gulf of Martaban. Mar. Geol.; 2004; 208, pp. 33-42. [DOI: https://dx.doi.org/10.1016/j.margeo.2004.04.019]
22. Li, Y.; Li, J. Suspended sediment remote sensing algorithm based on the phenomenon of spectral reflectance slope transfer between sea surface and remote sensor. Chin. Sci. Bull.; 1999; 17, pp. 1892-1897.
23. Liu, D.Z.; Zhang, C.G.; Fu, D.Y.; Shen, C. Hyperspectral data-based remote- sensing inversion model for suspended sediment in surface waters at the Pearl River Estuary. Mar. Sci.; 2010; 34, pp. 77-80.
24. Kazemzadeh, M.B.; Ayyoubzadeh, S.A.; Moridnezhad, A. Remote Sensing of Temporal and Spatial Variations of Suspended Sediment Concentration in Bahmanshir Estuary, Iran. Indian J. Sci. Technol.; 2013; 6, pp. 5036-5045. [DOI: https://dx.doi.org/10.17485/ijst/2013/v6i8.10]
25. Montanher, O.C.; Novo, E.M.; Barbosa, C.C.; Rennó, C.D.; Silva, T.S.F. Empirical models for estimating the suspended sediment concentration in Amazonian White water rivers using Landsat5/TM. Int. J. Appl. Earth Obs. Geo Inf.; 2014; 29, pp. 67-77.
26. Ma, W.D.; Wu, C.Q.; Li, L.; Wang, Y.M. Remote Sensing Retrieval of Total Suspended Matter in Hangzhou Bay Based on HJ-CCD. Proceedings of the 16th China Environmental Remote Sensing Application Technology Forum; Nanning, China, 28 March 2012.
27. Li, X. An United Equation for Remote Sensing Quantitative Analysis of Suspended Sediment and Its Application at the Pearl River Estuary. Remote Sens. Environ.; 1992; 7, pp. 106-114.
28. Han, B.; Loisel, H.; Vantrepotte, V.; Mériaux, X.; Bryère, P.; Ouillon, S.; Dessailly, D.; Xing, Q.; Zhu, J. Development of a Semi-Analytical Algorithm for the Retrieval of Suspended Particulate Matter from Remote Sensing over Clear to Very Turbid Waters. Remote Sens.; 2016; 8, 211. [DOI: https://dx.doi.org/10.3390/rs8030211]
29. Liu, X.P.; Deng, R.R.; Peng, X.J. An Integrated Model for Quantitative Remote Sensing Measurement of Suspended Sediment and Its Application in the Pearl River Estuary. Acta Sci. Nat. Univ. Sunyatseni; 2005; 44, pp. 109-113.
30. Deng, R.R.; He, Z.J.; Cheng, X.X. Model for water pollution remote sensing based on double scattering and its application on the the Pearl River Estuary. Acta Oceanol. Sin.; 2003; 25, pp. 69-78.
31. Liu, F.F.; Chen, C.Q.; Tang, S.L.; Liu, D.Z. A piecewise algorithm for retrieval of suspended sediment concentration based on in situ spectral data by MERIS in the Pearl River estuary. J. Trop. Oceanogr.; 2009; 28, pp. 9-14.
32. Xi, H.X.; Zhang, Y.Z. Total suspended matter observation in the Pearl River Estuary from in situ and MERIS data. Environ. Monit. Assess.; 2011; 177, pp. 563-574. [DOI: https://dx.doi.org/10.1007/s10661-010-1657-3]
33. Zhu, F.; Ou, S.Y.; Zhang, S.; Luo, K. MODIS images-based retrieval and analysis of spatial-temporal change of superficial suspended sediment concentration in the Pearl River Estuary. J. Sediment Res.; 2015; 2, pp. 67-73.
34. Luan, H.; Fu, D.Y.; Li, M.J.; Zeng, J.S.; Chen, Q.D. Based on Landsat 8 suspended sediment concentration of the Pearl River on each season inversion and analysis. Mar. Environ. Sci.; 2017; 36, pp. 892-897.
35. Liu, D.Z.; LI, Z.; Chen, Z.H.; Cheng, Y.C. Influence of Hong Kong-Zhuhai-Macao Bridge on the Distribution of Suspended Sediment in the Pearl River Estuary. Mar. Environ. Sci.; 2020; 40, pp. 90-96.
36. Jia, G.; Xu, S.; Chen, W.; Lei, F.; Bai, Y.; Huh, C. 100-year ecosystem history elucidated from inner shelf sediments off the Pearl River Estuary, China. Mar. Chem.; 2013; 151, pp. 47-55. [DOI: https://dx.doi.org/10.1016/j.marchem.2013.02.005]
37. Wei, X.; Wu, C.Y. The formation and development of the deposition bodies and main channel in the Zhujiang River Delta. Haiyang Xuebao; 2018; 40, pp. 66-78.
38. Liu, C.J.; Xia, H.Y.; Wang, D.Y. The observation and analysis of eastern Guangdong coastal downwelling in the winter of 2006. Acta Oceanol. Sin.; 2010; 32, pp. 1-9.
39. Xiao, Z.J. Characteristics and transport trend of surface sediments in Pearl River Estuary and the adjacent sea area. Mar. Sci. Bull.; 2012; 31, pp. 481-488.
40. Zhan, W.K.; Wu, J.; Wei, X.; Tang, S.; Zhan, H. Quantile trend analysis for suspended sediment concentration in the Pearl River Estuary based on remote sensing. J. Trop. Oceanogr.; 2019; 38, pp. 32-42.
41. Gao, F.; Wang, Y.P.; Hu, X.Y. Experimental Research on the Selection of Coagulant for Slightly Polluted Cellar Rainwater Treatment. J. Green Sci. Technol.; 2018; 6, pp. 1-5.
42. Pozdnyakov, D.; Grassl, H. Color of Inland and Coastal Water; Springer: Berlin/Heidelberg, Germany, 2003; pp. 81-120.
43. Zhou, Y.; Zhou, W.Q.; Wang, S.X.; Zhang, P. Applications of remote sensing techniques to inland water quality monitoring. Adv. Water Sci.; 2004; 15, pp. 312-317.
44. Zhang, K.; Zhang, K.; Niu, P.T.; Gao, L. Research Progress of Water Quality Monitoring Technique Based on Remote Sensing. Mod. Min.; 2018; 34, pp. 171-174+202.
45. Gordon, H.R.; Clark, D.K. Clear water radiance for atmospheric correction of Coastal Zone Color Scanner imagery. Appl. Opt.; 1981; 20, pp. 4175-5180. [DOI: https://dx.doi.org/10.1364/AO.20.004175] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20372349]
46. Shu, X.Z.; Yin, Q.; Kuang, D.B. Relationship between Algal Chlorophyll Concentration and Spectral Reflectance of Inland Water. J. Remote Sens.; 2000; 4, pp. 41-45.
47. Härmä, P.; Vepsäläinen, J.; Hannonen, T.; Pyhälahti, T.; Kämäri, J.; Kallio, K.; Eloheimo, K.; Koponen, S. Detection of water quality using simulated satellite data and semi-empirical algorithms in Finland. Sci. Total Environ.; 2001; 268, pp. 107-121. [DOI: https://dx.doi.org/10.1016/S0048-9697(00)00688-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11315735]
48. Thiemann, S.; Kaufmann, H. Lake water quality monitoring using hyperspectral airborne data—A semiempirical multisensor and multitemporal approach for the Mecklenburg Lake District, Germany. Remote Sens. Environ.; 2002; 81, pp. 228-237. [DOI: https://dx.doi.org/10.1016/S0034-4257(01)00345-5]
49. Zhang, Y.; Xia, H.Y.; Qian, L.B.; Zhu, P.L. Analysis on hydrological characteristics off the Pearl River Estuary in summer and winter of 2006. J. Trop. Oceanogr.; 2011; 30, pp. 20-28.
50. Han, W.Y.; Ma, K.M. A study on the Upwelling along the coast of eastern Guangdong. Acta Oceanol. Sin.; 1988; 10, pp. 52-59.
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Liu, Yingying 4 ; Tsou, Jin Yeu 5
; Zhang, Yuanzhi 6 1 Zaozhuang Meteorological Bureau, Zaozhuang 277100, China;
2 Zaozhuang Meteorological Bureau, Zaozhuang 277100, China;
3 South China Institute of Environmental Sciences, Ministry of Ecology and Environment of PRC, Guangzhou 510655, China;
4 School of Marine Sciences, Nanjing University of Information Science and Technology, Nanjing 210044, China;
5 Department of Architecture and Civil Engineering, City University of Hong Kong, Hong Kong 999077, China;
6 School of Marine Sciences, Nanjing University of Information Science and Technology, Nanjing 210044, China;