Introduction

Turbidity, along with other variables related to light penetration or suspended matter, has been the subject of scientific attention since the beginnings of classical studies of the marine environment, such as those of Le Grand1, among others cited in Wilber2. Turbidity has always been an important parameter across various disciplines, including biology, geology, and engineering. It serves as a key indicator in assessing the health of estuarine habitats due to its connection with various ecosystem components and its relevance to environmental management3,4. When its values are too high, turbidity affects phytoplankton and macroalgae, decreasing primary production by preventing light penetration5. It also affects fauna, especially filter feeders, but in this case, its effect can be negative as well as positive. Too low turbidity values may be related to food availability, as there is a relationship between the filtration rate and the concentration of particles in the water6,7. On the other hand, too high turbidity can cause gill damage or be associated with high sedimentation rates that alter sediment characteristics, either of which can cause economic damage to estuarine shellfish beds8. In addition, turbidity influences predator-prey relationships by reducing the efficiency of visual predators and providing a competitive advantage to chemosensory predators9. Furthermore, turbidity may modify the faunal composition and diversity of benthic communities10.

Turbidity depends on both the size and quantity of suspended particles. Suspended particles can be either inorganic or organic suspended sediment3 or originating mainly from planktonic organisms11,12,13 and their relationship with bacterial activity14,15,16. Consequently, turbidity is closely linked to sediment resuspension, sewage discharge, riverine inputs and its own turbidity, and plankton abundance17,18. Furthermore, bacterial activity, diazotroph organisms, macrophytes and microphytobenthos are also relevant in certain systems13. Therefore, in some ecosystems like rias, turbidity is mainly associated with: (i) wind, due to its impact on sediment resuspension10,14 and upwelling, which induces variations in nutrients concentration19; (ii) solar irradiation in relation with the photogeochemestry of particulate organic matter20 and phytoplankton biomass21,22; (iii) rainfall, leading to fluctuations in river flow, salinity, and sediment and wastewater input23; and (iv) tidal cycles, given their potential to induce sediment resuspension in specific locations, or their influence on the position of the point of maximum turbidity in estuaries24,25,26. However, the relative contribution of each parameter influencing turbidity varies depending on the specific oceanographic characteristics of each estuary13, as well as the quantity and quality of wastewater inputs. In addition, the oceanographic characteristics can be altered. Human activities can alter the hydrodynamics of the estuary, the wastewater input or the coastline. Therefore, medium to long-term changes in estuarine characteristics or inland inputs may be reflected in turbidity patterns, suggesting that turbidity may be an effective indicator of such changes.

Study area

The ria of Ferrol (Galicia - NW Spain) has a length of 15 km and a maximum depth of 33 m in its outer zone. This ria is narrower than the others in Galicia (2.3 km maximum width in its inner zone) and narrows as it approaches the sea. The Ferrol strait (0.5 km wide), located in the outer zone (Fig. 1), restricts water exchange with the open ocean27. The hydrography of this ria has been extensively studied in the first decade of the 2000s14,27,28,29. Thermohaline stratification induced by continental runoff is usually present in the inner part of the ria, and upwelling of shelf water affecting the water layer below 15 m is observed its outer part during summer14. Due to the geographical position and orientation of the ria, water renewal from the shelf is primarily driven by northerly winds21,28, although tides exert the most significant influence on circulation27. Due to the Ferrol strait, nutrient dynamics are also strongly influenced by the resuspension of sediments accumulated in the interior of the ria27,28. Mean precipitation in the Xubia river basin, which discharges into the inner part of the ria, ranged from 137 l/m² in summer to 521 l/m² in autumn between 2005 and 2023 (www.meteogalicia.gal). Although the Xubia river exhibits a relatively low river flow range (2 m3/s in summer to 18 m3/s in winter 2005 and 2023; www.meteogalicia.gal), its influence on the ria’s water mass is greater than observed in other rias in Galicia27.

In the latter half of the 20th century, the Ferrol ria experienced significant changes in the surrounding population, industrial development, and land use30. However, the first two decades of the 21 st century witnessed a series of notable meteorological and land-use changes that require careful consideration in any study of the ria. In 2003, the construction of a harbour breakwater outside the ria reduced the width of its mouth from 2.4 km to 1.4 km (Fig. 1). In 2007, the shellfish beds in the inner part of the ria were classified as Zone C (prohibited for bivalve mollusc exploitation due to bacterial contamination), according to European regulations31, due to high bacterial loads from urban wastewater. Between 2014 and 2017 upgrades to the sewage network and the connection of wastewater discharges to treatment plants resulted in a 97% reduction in untreated wastewater entering the ria by 201732. These improvements contributed to a reduction in bacterial loads, enabling the reclassification of the shellfish beds within the ria as Zone B (allowing the marketing of bivalve molluscs following purification) in 2018, as per Regulation 853/2004 (EC)33. On the other hand, while winds from the N (0º) and NE (45º) dominate in spring and summer, and those from the SW (225º) in winter, a shift in the wind regime has been observed. Between 2015 and 2023 an absence of days with a dominance of N (0º) and NW (315º) winds has been registered at the CIS-Ferrol weather station, located at 37 m altitude in the interior of the ria (Fig. 1). In the meantime, the number of days dominated by S (180º) winds has increased (Fig. 2). N (0º) and NW (315º) winds favour the renewal of water and nutrients in the ria via upwelling28. Nevertheless, assessing the influence of these changes on the environmental conditions within the ria presents a significant challenge. However, potential impacts on seawater turbidity can serve as an effective indicator of the ecosystem’s vulnerability to these environmental perturbations.

The primary objective of this study is to identify the key variables that exert the most substantial influence on turbidity within the ria of Ferrol. The results of this study may help to identify the variables that make the system vulnerable to climate change and anthropogenic interventions, particularly considering the recent observed changes in meteorological conditions, coastal line and wastewater management.

Fig. 1
figure 1

Study area. The meteorological stations of CIS-Ferrol and Aldea Nova; the turbidity monitoring station (AP Station); the location for collecting water samples for nutrient concentration monitoring (Barallobre) and the wastewater treatment plants (WTP) in the ria of Ferrol are highlighted. The numbers indicate depth in meters. The coordinates are expressed in UTM; DATUM WGS84. Created with QGIS v 3.28 based on the spatial data infrastructure of the Hydrographic Institute of the Spanish Navy available for free use at https://ideihm.covam.es.

Full size image
Fig. 2
figure 2

Number of days with wind dominance from each direction at CIS-Ferrol meteorological station.

Full size image

Materials and methods

To assess the potential factors contributing to turbidity in the inner ria of Ferrol, this variable was correlated with wind speed and direction, precipitation, tidal state, solar irradiation, and the phosphate and nitrate concentrations in the water.

Sampling

Turbidity was measured three to four times per week between January 2021 and February 2024 at the AP monitoring station overseeing the exploitation of the As Pías shellfish bed, located within the inner ria (Fig. 1). During high tide, a water sample was collected using a Niskin bottle at a depth of one meter above the seabed (3–6 m depth). Turbidity was then measured with a Turbiquant 118,325 Merck turbidity meter, recording values in nephelometric turbidity units (NTU). The reported turbidity values represent the average of ten measurements.

Data collection

Daily data on prevailing wind speed and direction, along with solar irradiation (10 kJ/m²day) were taken from the CIS-Ferrol weather station (www.meteogalicia.gal) (Fig. 1). The absence of N (0º) and NW (315º) wind data (Fig. 2) demanded the acquisition of wind direction data from an additional station in the vicinity of the ria to discard potential measurement errors at the CIS-Ferrol station. Consequently, data recorded at the Aldea Nova weather station were also consulted. The Aldea Nova weather station is located at 278 m altitude at the head of the valley that forms the ria, 5.5 km from the mouth of the Xubia river (Fig. 1). A similar pattern in wind dynamics observed at this station (Fig. 3) corroborated the validity of the data from the CIS-Ferrol station, which were ultimately used in the analysis. Daily rainfall data were obtained from the Aldea Nova weather station, considered representative of precipitation within the ria’s catchment area. Tidal data in the ria of Ferrol were also acquired from www.meteogalicia.gal, and subsequently, the maximum daily tidal amplitude was calculated. The Technological Institute for the Control of the Marine Environment of Galicia (INTECMAR) provided data on phosphate and nitrate concentrations (µmol/L of NO3− and µmol/L of PO43−) analysed in water samples collected weekly at a monitoring station located in Barallobre (Fig. 1) by the technical staff of the local fishermen’s association. The sample collection has received the necessary permissions from the competent bodies.

Fig. 3
figure 3

Number of days with wind dominance from each direction at Aldea Nova weather station. * The station’s record begins in 2009.

Full size image

Data processing

To align the temporal scales of the different variables, daily precipitation and solar irradiation values were transformed into accumulated data, considering the previous six days before each turbidity measurement. Following Alcantara et al.34, the daily wind direction and wind speed data for each day of turbidity measurements were used to derive two new variables: resuspension wind speed (RWS) and non-resuspension wind speed (NRWS). RWS was defined as the wind speed above the minimum threshold required to initiate sediment resuspension at low tide prior to each day’s turbidity measurement, while NRWS included wind speeds below this threshold. These variables were calculated for each wind direction (0º, 45º, 90º, 135º, 180º, 225º, 270º, 315º). The minimum wind speed (Uc) necessary to generate waves capable of reaching the seabed and initiating sediment resuspension at the first low tide of each day was determined as

$$\:Uc=1.2\:{\left\{4127\frac{{Tc}^{3}}{F}\right\}}^{0.813};\:\:\:\:\:\:Tc={\left(\frac{4\pi\:d}{g}\right)}^{\frac{1}{2}}\:\:\:\:$$

35 where Tc is the critical wave period36; d is the water depth corrected for the tidal height at the AP turbidity measurement station (Fig. 1); g is the acceleration due to gravity and F is the fetch, the space for the wind to generate the wave. To calculate F, the method described by Carper and Bachmann37, was used. Seven radials were plotted at 6° intervals (three on each side of the main radial for each wind direction), with the centre of the As Pías bridge near the AP station as the origin. (Fig. 4). The fetch (F) for each wind direction was then calculated by summing the lengths of these seven radials as they intersected the coastline and applying the following equation:

$$\:F=\frac{{\sum\:}_{i=1}^{i=7}{\left(\text{cos}{\alpha\:}_{i}\right)}^{2}{F{\prime\:}}_{i}}{{\sum\:}_{i=1}^{i=7}{\left(\text{cos}{\alpha\:}_{i}\right)}^{2}}$$

where α is the direction of each of the seven radials in radians and F’ is the fetch associated with each individual radial for the respective wind direction37.

Fig. 4
figure 4

Diagram for the calculation of the fetch (F) in the centre of the As Pías bridge (Ferrol). Created with QGIS v 3.28 based on the spatial data infrastructure of the Hydrographic Institute of the Spanish Navy available for free use at https://ideihm.covam.es.

Full size image

All variables were smoothed using LOESS38, employing a moving regression with 90 data points and a smoothing parameter of α = 0.148. This moving regression was chosen because it is approximately one quarter and its corresponding smoothing parameter (< 0.25) is considered weak and appropriate to preserve the seasonality of the data, if any. After smoothing, Pearson’s correlation analysis was conducted for each variable in relation to turbidity. Subsequently, to account for the disparities in the ranges of values and units across the different variables, the data were standardised according to the following equation:

$$\:{x}^{{\prime\:}}=\:\frac{x-\mu\:}{\sigma\:}$$

where µ represents the mean of the values (x) for each variable, and σ represent its standard deviation. Following the verification of relationships between variables through correlation analysis, those variables not correlated with turbidity were excluded. Then, a multiple linear regression (MLR) model was fitted using the smoothed data to investigate the dependence of turbidity on all remaining variables. To identify and eliminate any redundant variables, the Variance Inflation Factor (VIF) statistic was calculated for each independent variable:

$$\:{VIF}_{x1}=\frac{1}{1-{r}_{xk-1}^{2}}$$

where x1 is independent variable 1 and r²xk−1 is the correlation coefficient of x1 with respect to the other independent variables39. The results of the MLR model were compared with findings from other ongoing studies investigating changes observed within the ria, such as the relationship between N (0º) and NW (315º) winds and phosphate and nitrate concentrations since 2005, as well as sediment characteristics. In June 2013 and April 2024, the organic matter content of the sediment was analysed at 31 monitoring stations situated on the As Pías shellfish bed, near the AP station (Fig. 1), where turbidity data were collected. Organic matter content was determined as the percentage of weight loss on calcination of the particle size fraction smaller than 0.5 mm at 450 °C for 4 h40.

Results

Relationship between turbidity and the studied variables

Turbidity values ranged from 0.01 NTU in December 2021 to 12.93 NTU in April 2022, exhibiting pronounced seasonality with maxima observed during the spring and summer months (Fig. 5). During the study period, winds from the N (0º), E (90º), SE (135º) and NW (315º) did not reach sufficient speed to cause sediment resuspension. For winds from S (180º) and W (270º), resuspension events were recorded on only 25 and 16 days, respectively, out of the 609 days studied. Consequently, only NE (45º) and SW (225º) winds were considered for variables related to resuspension-inducing wind speed (RWS). Conversely, N (0°), SE (135°) and NW (315°) winds were excluded from the non-resuspension wind speed (NRWS) variables because they were recorded at non-resuspension speeds on less than 50 of the 609 study days. So, only winds from the NE (45º), E (90º), S (180º), SW (225º), and W (270º) were considered for variables related to non-resuspension wind speed (NRWS). Precipitation and SW (225º) wind speeds during resuspension events exhibited the same trends, generally displaying an inverse relationship with NE (45º) wind speed during resuspension (Fig. 5). The variation of wind speed for different wind directions when resuspension was not occurring is illustrated in Fig. 6. Overall, turbidity fluctuations were directly related to irradiation, with high values observed during spring and summer and lower values during autumn and winter. Furthermore, turbidity variations were inversely associated with phosphate and nitrate concentrations (Fig. 7).

Fig. 5
figure 5

Variation of the LOESS smoothed and normalised values of turbidity, precipitation, NE (RWS45) and SW (RWS225) wind speed when causing resuspension, and daily maximum tidal range (Max. tidal range).

Full size image
Fig. 6
figure 6

Variation of the LOESS smoothed and normalised values of turbidity and wind speed of NE (45º) (NRWS45), E (90º) (NRWS90), S (180º) (NRWS180) and W (270º) (NRWS270) when not causing resuspension.

Full size image
Fig. 7
figure 7

Variation of the LOESS smoothed and normalised values for turbidity, phosphates and nitrates content, and weekly irradiation.

Full size image

Turbidity was correlated with variables investigated, except for SW (225º) wind speed when it did not induce resuspension (Table 1). This variable was, therefore, excluded from the regression model. All variables included within the MLR model contributed to the regression with turbidity. The regression coefficients of the normalised variables showed that phosphate concentration (P), solar irradiation (WI), maximum tidal amplitude, and precipitation were the most influential variables on turbidity in the regression model, in that order. Among wind-related variables, NE (45º) winds contributed most to turbidity during non-resuspension periods, while SW (225º) winds were the primary driver during resuspension events. The MLR model proved to explain 93% of the observed variation in turbidity (r²adj = 0.9304, n = 609, p < 0.0001) (Table 2). The model variables showed no evidence of multicollinearity (VIF < 10) and were fitted to the following equation:

$$\begin{aligned}NTU &= 0.055 \cdot PP-0.026 \cdot {RWS}_{45}-0.047 \cdot {RWS}_{225}-0.121 \cdot {NRWS}_{45}\:+0.02{\cdot NRWS}_{90} + 0.011 \cdot {NRWS}_{180} \\&+ 0.072 \cdot {NRWS}_{270}-1.726 \cdot MaxTide-2.75 \cdot P-0.016 \cdot N+0.0004\cdot WI + 9.06\end{aligned}$$

where NTU represents turbidity; PP denotes precipitation; RWS45 and RWS225 represent the NE (45º) and SW (225º) wind speeds during resuspension events; NRWS45, NRWS90, NRWS180 and NRWS270 are the NE (45º), E (90º), S (180º) and W (270º) wind speeds during non-resuspension periods; MaxTide represents the maximum tidal amplitude; P and N denote phosphate and nitrate concentrations in the water, respectively; and WI is the weekly solar irradiation.

Table 1 Pearson’s correlation coefficient (r) between turbidity and the rest of variables after LOESS transformation. PP, precipitation; RWS45 and RWS225, NE (45º) and SW (225º) wind speeds that generated resuspension; NRWS45, NRWS90, NRWS180, NRWS225 and NRWS270, NE (45º), E (90º), S (180º), SW (225º) and W (270º) wind speeds that did not generate resuspension; P, phosphate content; N, nitrate content; WI, weekly irradiation; MaxTide, maximum daily tidal amplitude. Sample size = 609.
Full size table
Table 2 Results of the multiple linear regression model. PP, precipitation; RWS45 and RWS225, NE (45º) and SW (225º) wind speeds that generated resuspension; NRWS45, NRWS90, NRWS180, NRWS225 and NRWS270, NE (45º), E (90º), S (180º), SW (225º) and W (270º) wind speeds that did not generate resuspension; P, phosphate content; N, nitrate content; WI, weekly irradiation; MaxTide, maximum daily tidal amplitude. SE, standard error of the coefficient.
Full size table

Relationship among the observed changes in the Ria

The mean annual phosphate concentration in the internal waters of the ria has decreased from 0.67 µmol/L in 2005 to 0.3 µmol/L in 2023. Nitrate concentrations showed the highest value in 2006 (17.388 µmol/L) and the lowest values in 2017 (5.4 µmol/L) and 2022 (6.092 µmol/L). Phosphate concentration changed with the annual number of NW (315º) and N (0º) wind days, with Pearson´s correlation coefficient (r) of 0.598 (p > 0.01, n = 18) and 0.609 (p > 0.01, n = 18), respectively, based on an 18-year dataset (Fig. 8). Mean annual nitrate concentration did not exhibit correlation with the number of NW (315º) and N (0º) wind days, although it was directly correlated to phosphate concentrations (r = 0.519, p > 0.05, n = 18). Concurrently, a decrease in sediment organic matter content was observed between 2013 and 2024. At 21 out of the 31 monitoring stations located on the As Pías shellfish bed, near the AP station, a decline in organic matter content was evident. The range of sediment organic matter content at these stations in 2013 varied from 7.7% to 1.4% (x̅ = 4.1%), while in 2024, the range was from 5.1% to 0.6% (mean = 2.5%). The observed reductions in organic matter content at the different stations ranged from 75% to 11%.

Fig. 8
figure 8

Relationship of phosphate and nitrate content with observed changes in wind regime (days/year of NW (315º) and N (0º) winds) and organic matter (O.M.) content in the sediment. The period when the inner part of the ria was declared as zone C, due to excess E. coli derived from sewage, and the period when the sewage was connected to sewage treatment plants are noted.

Full size image

Discussion

The regression model fitted to the studied variables demonstrated that, except for SW (225°) wind speeds insufficient to induce resuspension (NRWS225), all considered variables contributed to the variation of turbidity. These variables included: NE (45°) and SW (225°) wind speeds during resuspension events (RWS45 and RWS225); NE (45°), E (90°), S (180°), and W (270°) wind speeds during non-resuspension periods (NRWS45, NRWS90, NRWS180 and NRWS270); riverine inputs (represented by precipitation, PP); tidal amplitude (MaxTide); nutrient concentrations in the water (P and N); and solar irradiation (WI). This model accounted for more than 93% of the observed variation in turbidity (Table 2).

Importantly, the regression coefficients revealed that precipitation and winds from E (90º), S (180º) and W (270º) with speeds insufficient to induce resuspension exhibited positive relationships with turbidity. Precipitation increases turbidity by enhancing the input of suspended matter from riverine sources, while low-velocity winds from the E (90º), S (180º) and W (270º) contributed to the accumulation of suspended matter within the ria due to its specific geographical configuration (Table 2). Conversely, low-velocity NE (45º) winds exhibited a negative relationship with turbidity in the model. The direction of these winds aligns with the geographical orientation of the study area, facilitating the ‘washing out’ of suspended matter from the interior of the ria to the ocean. Similarly, tidal amplitude also exhibited a negative relationship with turbidity, consistent with the observed inverse oscillation between these two variables (Fig. 5). This inverse relationship may be attributed to the influence of tides on circulation within the ria29 or to the impact of tidal fluctuations on the proximity of the zone of maximum turbidity to the AP sampling station, given the influence of tides on the displacement of this zone24.

As illustrated in Fig. 7, turbidity varied concurrently with solar irradiation, and both was inversely correlated with phosphate and nitrate concentrations. Increased concentrations of these nutrients were observed during the autumn and winter months, which coincide with periods of reduced solar irradiation. This temporal variation in nutrient concentrations has been previously described within the Ferrol ria by Bode et al.28. Similarly, Nielsen et al.19 related the concentration of chlorophyll and these nutrients with suspended matter. Furthermore, they found that the relationship between chlorophyll and suspended matter was stronger in summer, when nutrient concentrations were lower. This suggests that increased phytoplankton biomass, driven by enhanced solar irradiation and the consumption of these nutrients, contributes to the observed increase in turbidity during spring and summer. Given the limited influence of upwelling within the ria of Ferrol, attributed to the restricted water exchange between the ria and the open ocean due to the presence of the Ferrol Strait27,41, nutrient replenishment within the ria is primarily driven by internal remineralisation processes. So, nutrient concentrations are likely enhanced by resuspension events triggered by NE (45°) and SW (225°) winds. This explains the negative contribution of the speed of these winds to turbidity during resuspension events in the regression model (Table 2). The influence of phytoplankton on turbidity or suspended matter has been previously documented by Nielsen et al.19, just as Bailey and Hamilton11 indicate that phytoplankton interfere with suspended solids measurements. Bacterial activity and its complex relationship with phytoplankton and particulate organic matter also play an important role15,16. Understanding the sensitivity of this relationship to anthropogenic change is a challenge16. However, its influence on total suspended matter is clear. The relationship between bacterial activity and phytoplankton can be very different depending on the type of estuary [16 and others cited by them]. In the case of the ria de Ferrol, Varela et al.14 found that the maximum of bacteria occurs with some delay to that of phytoplankton, so that bacteria grow throughout the summer taking advantage of the substrate produced by phytoplankton, as also found by Crump et al.15. Thus, the influence of phytoplankton blooms and bacterial activity on turbidity may persist throughout the summer, as shown in Fig. 7.

In this study turbidity was related to rainfall, wind speed of different origins, nutrient content in the water, solar irradiation and tidal amplitude. These variables can be grouped into two components, one hydrodynamic and the other biological. The hydrodynamic component of the studied variables includes: (i) river flow, (ii) tidal amplitude, and (iii) wind speed. Rainfall-driven river flows can resuspend sediments and introduce suspended matter in the estuaries4,34. Tidal fluctuations, together with the entire hydrodynamic configuration and gravitational circulation, influence turbidity23,25,42. Winds aligned with the ria’s orientation generate waves that resuspend sediments when reaching sufficient speeds11,34. Strong winds can also “wash out” suspended matter from the interior of the ria to the ocean. This hydrodynamic component contributes to turbidity variations, in agreement with findings from other studies4,18. The biological component is linked to phytoplankton dynamics and bacterial activity11,16,19. Other groups such as macrophytes or microphytobenthos may also be related to particulate organic matter13. This study indirectly assessed this relationship through correlations with solar irradiation and phosphate and nitrate concentrations19,22. The biological component depends on the nutrients input to the ria, whether through upwelling or internal remineralization, with photochemical transformations probably playing a critical role in the bioavailability of organic matter20. While NE (45°) winds can induce upwelling outside the ria27,43 suggest that upwelling has a limited impact within the inner ria. Consequently, nutrient variations are primarily driven by internal remineralization processes, influenced by wind and tidal mixing28,29. In fact, this study observed a strong correlation between phosphate concentration and northerly (NW and N) winds (Fig. 8). Regression model revealed that the variables that contributed most to turbidity were phosphate concentration and solar irradiation (biological components) followed by tidal amplitude and precipitation (hydrodynamic components) (Table 2).

Both hydrodynamic and biological components contributing to turbidity in the ria have changed significantly over the last two decades. The wind regime has shifted, with the notable lack of N and NW winds between 2015 and 2023. Although no direct relationship with turbidity was observed during the study period due to the absence of these winds, historical data suggests that northerly winds previously caused sediment resuspension and likely enhanced phosphate concentrations11. This is supported by a decline in phosphate levels observed since the cessation of these winds (Fig. 8). In addition, a decrease in organic loading from sewage is evident. The closure of the As Pías shellfish bed in 2007 due to high E. coli levels, indicative of urban wastewater discharges, highlights this issue. Subsequent upgrades to the sewerage network and wastewater treatment plants between 2014 and 2017 led to improved water quality and reclassification of the shellfish bed as zone B in 2018 (Fig. 8). A reduction in organic pollution is likely linked to decreased turbidity44 and may also contribute to the observed decline in phosphate levels45. Furthermore, Prego41 documented a strong association between phytoplankton dynamics and nutrient inputs from wastewater in the ria of Vigo. This is consistent with the observed decrease in organic matter content in the sediments of the inner ria of Ferrol between 2013 and 2024 (Fig. 8). This sediment modification is confirmed by changes in the benthic macrofauna, such as the recent appearance of Echinocardium cordatum (Pennant, 1777) specimens since 2020 in the fauna monitoring of the As Pías shellfish bed carried out since 2005 (Personal observation; unpublished data). The absence of E. cordatum is typically associated with organic pollution46 and its presence from 2020 onwards indicates a reduction in sediment organic matter content.

Turbidity is a valuable integrative indicator of ecosystem health within the ria. Changes in bacterial loads and benthic macrofauna communities point to a decrease in organic matter inputs from sewage sources, likely linked to improved wastewater treatment. These changes are reflected in reduced turbidity values. Modifications to the wind regime, specifically the lack of N and NW winds since 2015, also impact turbidity. These winds historically contributed to sediment resuspension and likely enhanced nutrient availability through both internal remineralization29 and upwelling from the open sea. However, the construction of the outer harbour in 2003 may have reduced the ria-ocean exchange, potentially limiting upwelling in this already relatively enclosed system27. These changes, caused by both human activities (wastewater treatment, harbour construction) and climate-related shifts in wind patterns, highlight the vulnerability of estuarine ecosystems to environmental perturbations. The above changes imply modifications in the relative weight of each of the variables influencing turbidity. Because of these significant environmental changes, the current regression model should not be applied directly to past conditions. However, the available evidence suggests a recent decrease in turbidity, contrary to the general trend of increasing turbidity in other estuaries10. Among the changes suggesting this decrease in turbidity, two are related to civil works and use of the environment: the improvement of the sewage treatment network and the construction of the harbour which reduced the width of the mouth by just over half (Fig. 1). A third change in the environmental conditions of the ria is related to the modification of the wind regime reflected in the disappearance of the N and NW winds. Historical data (Fig. 243; indicates that N and NW winds were more prevalent in the past. Although turbidity data are not available prior to the study, these winds likely contributed to increased turbidity through sediment resuspension and upwelling21. French et al.47 predicted that climate change could alter prevailing wind patterns, impacting sediment dynamics in estuaries. Our results support this prediction and demonstrate the vulnerability of sediment processes, nutrient balance and productivity of estuaries and rias to changes in the territory and climate, already observed in the wind regime. Although the study has been carried out in a ria, the results are applicable to estuarine systems in general, insofar as they are under the influence of fluvial dynamics and their circulation is modified by the exchange with oceanic waters. In fact, some of the works consulted [4, 5, 18, 23, 24, 42, 47, among others], describe similar relationships between the variables included in this study in different estuarine systems.

Conclusion

The results show that turbidity acts as an effective integrator of different environmental factors. By analysing turbidity and its associated variables, we gain valuable insights into the functioning of rias and their vulnerability to both meteorological and anthropogenic influences. The observed changes in (i) the wind regime; (ii) the improvement of the wastewater treatment and (iii) the narrowing of the mouth of the ria by the outer harbour prevented the application of the regression model to infer turbidity at times prior to the study. However, the results obtained suggest that the current levels of turbidity in the interior of the ria of Ferrol have decreased since the second decade of the 2000s.