Trends of Nitrogen and Phosphorus in Surface Sediments of the Lagoons of the Northern Adriatic Sea

: The analysis of nutrient concentrations in surface sediments is a reliable tool for assessing the trophic status of a water body. Nitrogen and phosphorus concentrations are strongly related to the sediment characteristics but are mainly driven by anthropogenic impacts. The results of the determination of total nitrogen and total inorganic and organic phosphorus in surface sediments of the lagoons and ponds of the northwestern Adriatic Sea (Marano-Grado, Venice, Po Delta, Comacchio Valleys, Pialassa della Baiona) show the merit of this approach. Indeed, when previous data are available, the ratio between the actual and background values can provide useful information on the trophic changes that have occurred in the most recent times, and the results can also explain the conditions present in less studied environments. In this context, numerous studies performed in the Venice lagoon since the second half of the 20th century during different environmental scenarios provide mean concentration ranges and propose the main causes of changes. The results of single datasets available for the other lagoons fall into scenarios that occurred in the Venice lagoon. At present, the most eutrophic basins are Pialassa della Baiona, the Po Delta lagoons and ponds and the Comacchio valleys due to industrial efﬂuents, ﬁsh farming and clam harvesting, respectively, whereas the Venice lagoon is now experiencing environmental recovery.

The lagoon of Venice is the largest TWS in the Mediterranean Sea (549 km 2 ) and presents the ecological conditions recorded in most of these coastal environments [24]. Data on the nitrogen and phosphorus concentrations in the surface sediments of the whole lagoon have been available since the second half of the 20th century [25][26][27] and are strongly correlated with the ecological changes observed over time [2]. Until the 1990s, a progressive increase of nutrients in the superficial sediments was observed, especially with regard to phosphorus, mainly due to (i) the production of nutrients by industrial synthesis (phosphates and ammonia); (ii) the draining waters from an areas of ca. 1800 km 2 extensively cultivated with monocultures fertilized with high amounts of nutrients; (iii) the discharge of urban wastewater not yet fully treated from the historic center of Venice, the islands of the lagoon and the city of Mestre and its hinterland [28][29][30]. In addition, the nutrient increase in the water column and surface sediments [18,31] favored the massive production of nuisance macroalgae [21], which during their degradation, further increased the load of nutrients in the surface sediments. The same impacts, especially the increase in trophy with abnormal development of macroalgae nuisance, also affected other TWSs both in Italy and in many other countries worldwide [5,6].
Later, more extensive studies carried out in 2003 and 2011 showed a significant decrease of these eutrophic substances due to industrial activity's decline, the implementation of water treatment plants equipped with a third stage for nitrogen and phosphorus abatement, the banning of phosphorus from detergent formulations (4-5% of total weight) in 1989 [1] and sediment remobilization during the intense Manila clam harvesting [15]. In 2014 and 2018, two additional extensive nutrient mapping studies were carried out but the data were again unpublished. Since 1987, for the central basin of the Venice lagoon, the TWS most studied due to its greater number of anthropogenic impacts, a greater number of mapping datasets are available for both phosphorus and nitrogen during different environmental scenarios.
On the other hand, there is little information from other TWSs of the Northern Adriatic Sea, data on which are scarce and fragmentary. Indeed, only a few data are available for the lagoons of Grado-Marano (2007) [32], the Po Delta (2008), Pialassa della Baiona (2009) and the Comacchio Valleys (2009) [33]. However, these data, all collected by our research group, are very useful to understand the correlations of nitrogen and phosphorus with the parameters associated with different ecological conditions. This paper aimed to integrate the knowledge on the nitrogen and phosphorus concentrations in the TWS of the northern Adriatic Sea with particular reference to the Venice lagoon, of which a high number of data is available. The complete dataset allows us to analyze the interactions of nutrients with other sediment characteristics and the ecological conditions of these TWSs as a function of anthropogenic impacts and climate changes. The results allow to understand the roles of these elements as indicators of ecological changes and for forecasting future trends.

Description of Study Areas and Sampling Campaigns
The lagoons and ponds investigated in this paper are reported in Figure 1. From the north to the northwestern Adriatic Sea, they are: Marano-Grado lagoon, Venice lagoon, Po Delta lagoons and ponds, Comacchio valleys and Pialassa della Baiona lagoon. The stations sampled in 2003 in the Venice lagoon were selected to study the biomass distribution of macrophytes [34]. All the stations sampled in the Venice lagoon since 2011 and the stations sampled in the other lagoons were selected by the Italian Regional Agencies for Environmental Prevention and Protection. Sampling was carried out to determine the ecological status of TWSs in the framework of the European Water Framework Directive 2000/60/EC by using the biological element "Macrophytes". On this occasion, our research group also collected surface sediments to study some sediment characteristics and nutrient concentrations.

The Venice Lagoon
The lagoon of Venice (45 • 11 -45 • 34 N, 12 • 08 -12 • 38 E) (Figure 1) is the largest and the most studied TWS of the Mediterranean Sea. It is located in the northwestern Adriatic Sea and extends over an area of ca. 549 km 2 accounting for ca. 39% of the total Italian TWSs [24]. The waters have a mean depth of ca. 1.2 m, ranging from a few centimeters in the shallower areas to 15-20 m in the main channels. The water exchange with the sea is mainly driven by tidal action [35]-through the three wide (600-900 m) and deep (12-15 m) inlets of Lido, Malamocco and Chioggia-and accounts for ca. 60% of the total reservoir every 12 h [36], whereas the freshwater inputs (ca. 34.5 m 3 s −1 ) are negligible [37]. This allows a high grain-size difference between the sandy areas close to the inlets where water exchange (ca 19,000 m 3 s −1 ) occurs in a few hours, and the clay chocked areas where water renewal may take 40 days [38]. The lagoon is morphologically separated in three main basins-from the salt marshes of Burano and Torcello in the north and the Malamocco-Marghera canal in the south-showing very different morpho-ecological conditions, the central basin being the most polluted and most affected by anthropogenic impacts.

The Venice Lagoon
The lagoon of Venice (45°11′-45°34′ N, 12°08′-12°38′ E) ( Figure 1) is the largest and the most studied TWS of the Mediterranean Sea. It is located in the northwestern Adriatic Sea and extends over an area of ca. 549 km 2 accounting for ca. 39% of the total Italian TWSs [24]. The waters have a mean depth of ca. 1.2 m, ranging from a few centimeters in the shallower areas to 15-20 m in the main channels. The water exchange with the sea is mainly driven by tidal action [35]-through the three wide (600-900 m) and deep (12-15 m) inlets of Lido, Malamocco and Chioggia-and accounts for ca. 60% of the total reservoir every 12 h [36], whereas the freshwater inputs (ca. 34.5 m 3 s −1 ) are negligible [37]. This allows a high grain-size difference between the sandy areas close to the inlets where water exchange (ca 19,000 m 3 s− 1 ) occurs in a few hours, and the clay chocked areas where water renewal may take 40 days [38]. The lagoon is morphologically separated in three main basins-from the salt marshes of Burano and Torcello in the north and the Malamocco-Marghera canal in the south-showing very different morpho-ecological conditions, the central basin being the most polluted and most affected by anthropogenic impacts.

The Lagoon of Marano-Grado
The Marano-Grado lagoon (45 • 41 -45 • 46 N, 13 • 04 -12 • 27 E) ( Figure 1) has morphological features similar to those of the Venice lagoon with a surface of ca. 160 km 2 and a tidal difference of 65 cm [36]. It is subdivided in two main basins: The Marano basin, a wide shallow water body characterized by a few islands and tidal-marshes, and the Grado basin, which is rich in morphological reliefs and is shallower than the Marano basin [40]. The average freshwater inflow is ca. 98.5 m 3 s −1 [41], whereas water exchange with the sea occurs through three main inlets (Lignano, Porto Buso, Grado) and other smaller mouths. Fine sediments are on average greater at Grado than at Marano and are prevalently calcareous. Information on the nutrient content is scarce. Falace et al. [42] report data for the total nitrogen (N tot ) only, with the concentration ranging from 0.5 mg g −1 at the sea-inlets, to 1.5 mg g −1 at the internal edges. Samples of surface sediments for the analysis of N tot , P tot , P inorg and P org were collected at 20 stations evenly distributed throughout the lagoon in summer 2007.

Po Delta Lagoons and Ponds
The Po Delta (44 • 09 -45 • 47 N, 12 • 16 -12 • 32 E) ( Figure 1) has a water surface of ca. 204 km 2 and is sorted into many lagoons and ponds with depths ranging from 0.5 to 2.5 m, according to the water body. Salinity is the most variable parameter due to the many branches of the Po River. Hard substrata are rare and mainly represented by some stone embankments along the edges of the various basins and oyster beds scattered on their bottoms. Clam-farming and clam-fishing activities are very intense and cause high sediment resuspension with severe environmental consequences. The main primary producers of the Po Delta basins are phytoplankton in the areas affected by clam harvesting or free-floating nuisance macroalgae, especially Ulvaceae, Gracilariaceae and Solieriaceae, in the other areas with lower anthropogenic impact. Populations of the aquatic angiosperm Ruppia cirrhosa (Petagna) Grande that before the 1990s colonized some basins have now completely disappeared [43]. Samples of surface sediments were collected in 2008 at 20 stations (three at Caleri, two at Marinetta, two at Vallona, three at Barbamarco, three at Canarin, four at Scardovari and three at Goro) during late spring-early summer period.

Comacchio Valleys
The Comacchio valleys (44 • 55 -44 • 65 N, 12 • 10 -12 • 25 E) ( Figure 1) are a complex of choked shallow ponds with a surface of ca. 110 km 2 , connected with the sea by two small channels that do not allow for sufficient water renewal. They are affected by anthropogenic impacts like fish farming, especially eel farming, and receive freshwater from the Reno River that has shifted the ecological conditions from a good ecological status dominated by aquatic angiosperms and macroalgae of high ecological value, to a degraded and homogeneous basin dominated by phytoplankton and picocyanobacterial blooms of poorto-bad status [44][45][46]. Samples for nutrient analyses of surface sediments were collected at two stations, Donna Bona and Dosso Pugnalino, in the main basin during late spring or early summer 2009.

Pialassa Della Baiona
The Pialassa della Baiona lagoon (44 • 28 -44 • 31 N, 12 • 14 -12 • 16 E) ( Figure 1) is a shallow basin of ca. 11 km 2 characterized by a mean depth of 60 cm. It is furrowed by a network of herringbone canals, 1-4 m deep, which converge towards the Ravenna harbor connected with the Adriatic Sea. The freshwater inputs, coming from a network of small canals draining an intensively cultivated area, are negligible. The basin is colonized by rich populations of macroalgae, especially the non-indigenous species Agarophyton vermiculophyllum [43,47] (Ohmi) Gurgel, J.N. Norris et Fredericq. This species covers mostly the southern part of the lagoon with a biomass of 5-10 kg m −2 on a fresh weight (fwt) basis. It is a little studied basin characterized by poor ecological conditions [48]. Surface sediments were sampled at three stations-Chiaro Magni, Chiaro Risega and Vena del Largo-in 2009 in late spring to early summer. For this basin, no other samples are available.

Sediment Sampling
At each station, three subsamples of surface sediments (5 cm top layer) were sampled by a Plexiglas corer (i.d. 10 cm) and mixed together. One subsample (ca. 100 mL) was retained for nutrient (total nitrogen, total phosphorus, inorganic phosphorus, organic phosphorus) analyses and another (ca. 50 mL) for the determination of the sediment density and grain size. Both subsamples were stored at −20 • C until the laboratory analyses.

Nutrient Determination in Surface Sediments
Information on the concentrations of P tot and N tot in the sediments of the whole lagoon date back to the mid-20th century. The analyses were carried out in the first 30 cm in 1948-49 and 1968-73 and in the first 20 cm in 1983 and 1987-88. In those periods, sediments were collected with a bucket to study the benthic communities [49,50]. During the following years, sediments were collected with a plexiglass corer (i.d. 10 cm) retaining the first 20 cm [27] or 5 cm top layer (this paper). This latter thickness is the sediment most affected by the growth and degradation cycles of macrophytes.
In the laboratory, sediments were freeze-dried and pulverized using a sediment mill (Fritsch Pulverisette, Germany). The concentration of total nitrogen (N tot ) was measured in duplicate by a CHNS Analyzer (Vario-MICRO, Elementar CHNS by Elementar Italia S.r.l.) after an accurate sample powdering of ca. 0.3 g of sample. The standard used was "low level N-and S-contents" with N = 0.74%, art. no. 05 000 959.
Total phosphorus (P tot ) was determined after sample combustion in the muffle at 550 • C for at least 2 h of 0.3-0.4 g of sample. Subsequently, the residue thus obtained was suspended in 50 mL of 1 N HCl and sonicated for ca. 30 min. After allowing the sample to settle for at least 1 h, 0.5 mL of the supernatant were taken with a graduated gaschromatographic syringe and brought to exactly 10 mL using volumetric flasks for a final dilution of 1 L, with the result expressed directly in µM. At this point, the phosphorus concentration was determined by spectrophotometry by adding the mixed reagent and reading the absorbance at 885 nm after ca. 10-15 min [51]. Inorganic phosphorus was obtained with the same procedure used for P tot but without combustion at 550 • C. Organic phosphorus (P org ) was determined by difference. All samples were analyzed in duplicate and the analyses were replicated on two different days to obtain an accuracy > 95. Otherwise, the analyses were repeated until the coefficient of variation (standard deviation/mean) between two replicates was < 5%.

Map Preparation
Maps of nutrient distribution were prepared in six concentration ranges by using data collected in the surveys carried out in 2003, 2011, 2014 and 2018. Maps were obtained by Kriging methods using the software R, version 4.0.3, and SAGA, version 2.3.2. Models for each analysis were selected to minimize the prediction errors, and model performances were estimated using cross-validation procedures. Maps were then produced using the software QGIS, version 3.6.0 and integrated into one image using Corel Paintshop Pro X5, version 15.0.0.183 (Corel Corporation, 2012).

Sediment Characteristic Determination
Dry sediment density (g DWT cm −3 ) was determined in the laboratory by sediment desiccation at 110 • C in tared crucibles of 20-30 mL. The percentage of Fines (fraction < 63 µm) was obtained by wet sieving ca. 50 g of dried sediment throughout Endecotts sieves (ENCO Scientific Equipment, Spinea, Italy). All analyses were performed in duplicate.

Statistical Analyses
The analysis of variance (one-way ANOVA) allowed us to test differences of the sediment parameters in different monitoring periods. The differences were considered significant when p < 0.05. Prior to the analyses, the distribution of each variable was checked for normality and homogeneity of variance by the Kolmogorov-Smirnov test (p < 0.05).
Principal component analysis (PCA), determined using the same software, showed the multivariate patterns of the matrix of 162 cases (stations) and five variables (sediment density, fines, inorganic and organic phosphorus, total nitrogen). The significant loading was considered to be at p > 0.7. Moreover, the PCA analysis of the transposed matrix showed the affinity between stations and lagoons.

Nutrient Concentrations and Sediment Characteristics
The mean values of the dry density, the percentages of the fine fraction and the concentrations of nutrients in the surface sediments of the single and total TWSs are reported in Figure 2 and  On the whole, the average dry density of all the sampling sites was 0.92 g cm −3 on a dry weight (dwt) basis. The highest density was recorded in the Venice lagoon (0.96 g dwt cm −3 ), whereas Pialassa della Baiona and the Comacchio valleys showed the lowest values (0.59 g cm −3 ). The amount of fines on average was 70.2%, with the highest and the lowest Figure 2. Average values of sediment dry density (g cm −3 ), fines (fraction < 63 µm), phosphorus species (P tot , P inorg , P org ) and total nitrogen (N tot ) in the studied TWSs.
On the whole, the average dry density of all the sampling sites was 0.92 g cm −3 on a dry weight (dwt) basis. The highest density was recorded in the Venice lagoon (0.96 g dwt cm −3 ), whereas Pialassa della Baiona and the Comacchio valleys showed the lowest values (0.59 g cm −3 ). The amount of fines on average was 70.2%, with the highest and the lowest concentrations in the Pialassa della Baiona (87.5%) and Comacchio valleys, (57.0%), respectively. Pialassa della Baiona also showed the highest concentrations of N tot (3.09 mg g −1 ), P tot (717 µg g −1 ), P inorg (537 µg g −1 ) and P org (180 µg g −1 ). Total phosphorus showed the lowest value at Marano (334 µg g −1 ), where N tot was also particularly low (1.23 mg g −1 ), just higher than that found in the Venice lagoon (1.19 mg g −1 ).

Nutrient Variations in the Venice Lagoon
The nutrient concentrations recorded in the whole Venice lagoon by our research team were recorded during four surveys carried out in late spring to early summer 2003, 2011, 2014 and 2018 (Table 1). Both P tot and N tot showed no significant changes. Indeed, the mean values of P tot recorded in this period were almost the same, ranging from 403 to 412 µg g Total phosphorus did not show significant changes but only a spatial redistribution with the highest concentrations found in the same areas of N tot . Organic phosphorus showed greater changes. Indeed, the mean concentration recorded in 2003 was 79 ± 56 µg g −1 . This value increased to 91 ± 63 µg g −1 in 2011, only to decrease to 81 ± 54 µg g −1 in 2018. The highest changes were recorded in the eastern parts of the central and southern basins.
The total nitrogen displayed the greatest changes. In late spring to early summer 1987, during the period of luxuriant macroalgal growth, the mean Ntot value was 1.25 mg g −1 (Table 2, Figure 4). The mean value in the period of clam harvesting decreased to 0.34 mg g -1 in late spring to early summer 2011, then increased slightly in 2014 and 2018 (0.83 and 0.81 mg g −1 , respectively). The decrease in the period 1987-2011 was −72.8%, whereas between 1987 and 2018 it was −35.2% (Figure 4). The total phosphorus and Pinorg showed no significant changes (Table 2) The one-way ANOVA of the Ntot and Porg concentrations in surface sediments of the central Venice lagoon between the successive sampling periods showed a different behavior for the two nutrients (Table 3). Nitrogen decrease occurred between 1993 and 2011 and was significantly related to clam harvesting activities. In contrast, Porg changes were only significant in the period 1987-1993 when the macroalgal biomass of nuisance macroalgae decreased sharply.  The total phosphorus and P inorg showed no significant changes (Table 2). Instead, P org had similar variations to N tot . Indeed, P org decreased from 109 µg g −1 in 1987 to 53 µg g −1 in 2003, and increased to 66 and 64 µg g −1 in 2014 and 2018, respectively. The difference between 1987 and 2003 was −51.4%, whereas between 1987 and 2018 it was −41.3% (Figure 4).
The one-way ANOVA of the N tot and P org concentrations in surface sediments of the central Venice lagoon between the successive sampling periods showed a different behavior for the two nutrients (Table 3). Nitrogen decrease occurred between 1993 and 2011 and was significantly related to clam harvesting activities. In contrast, P org changes were only significant in the period 1987-1993 when the macroalgal biomass of nuisance macroalgae decreased sharply.

Statistical Analyses
The statistical analysis (non-parametric Spearman's coefficients) of nutrient concentrations and sediment characteristics recorded in all the considered TWSs showed a high number of positive or inverse correlations between the sediment dry density and grain-size (fines) and the nutrient concentrations (Table 4). Among them, particularly relevant was the negative correlation of sediment density with P org (r = −0.84).  Figure 5a. The first two components explained 70.8% of the total variance and dry density. Organic phosphorus and P inorg showed a loading > 0.7. In addition, PCA highlighted the affinity among parameters, showing a strong association of nutrients with fines as opposed to the sediment density.  Table 4. Spearman's non-parametric coefficients between phosphorus species, total nitrogen, fines and density of surface sediments of all the lagoons and ponds considered. In blue are significant values: p < 0.05 for r ≥ ± 0.19; in red are highly significant values p < 0.001 for r ≥ ± 0.41. The PCA of the transposed matrix applied to the same parameters and stations (Figure 5b

Discussion
The concentrations of nutrients in the surface sediments of the lagoons of the northwestern Adriatic Sea showed a close relationship with sediment characteristics such as the grain size and density. In fact, finer-sized sediments had a greater relative surface and a greater capacity to retain nutrients and pollutants. Vaze and Chew [52] found that almost all the TP and TN were attached to sediments with grain sizes between 11 and 150 µ m. Furthermore, their density was lower due to the presence of a greater quantity of interstitial water, which in turn contains, high concentrations of nutrients, especially ammonium [53].
In the case of the Venice lagoon, of which we have numerous data collected in different periods, significant temporal variations due to different anthropogenic impacts were also monitored.
In general, the highest sediment concentrations of Ntot, Ptot, Pinorg and Porg were recorded in Pialassa della Baiona, characterized by the highest percentage of fines (fraction < 63 µ m: 87.5%) and the lowest density (0.59 g dwt cm −3 ). The same low sediment density was also recorded in the Comacchio valleys. However, the Comacchio valleys showed a significantly lower percentage of fines (57.0%), which could explain the lower concentrations of nitrogen and phosphorus in the sediments of this basin. Moreover, the poor water renewal of the Comacchio valleys accounted for the bad ecological conditions of these large ponds recorded since the last two decades of the 20th century. Despite the fact that for this basin, previous data on nutrient concentrations are missing, the list of macrophytes reported by [54] and [55] highlights that before the 1970s, the Comacchio valleys were colonized by aquatic angiosperms such as Ruppia cirrhosa (Petagna) Grande and Zostera noltei Hornemann, and sensitive macroalgae. First of all, these included Lamprothamnium papulosum (Wallroth) J. Grove, a Chlorophycea even more sensitive than aquatic angiosperms and currently present only in some fishing valleys of the Venetian lagoon such as the Cavallino valley, which shows almost pristine ecological conditions [43]. The presence of that vegetation was linked to a low trophy and nutrient concentrations certainly lower than the current ones. Indeed, in the following years, almost all the macrophytes disappeared from the valleys and only cyanobacteria and phytoplankton were recorded [44][45][46]. Discharges from fish farms, especially from eel breeding, the polluted water

Discussion
The concentrations of nutrients in the surface sediments of the lagoons of the northwestern Adriatic Sea showed a close relationship with sediment characteristics such as the grain size and density. In fact, finer-sized sediments had a greater relative surface and a greater capacity to retain nutrients and pollutants. Vaze and Chew [52] found that almost all the TP and TN were attached to sediments with grain sizes between 11 and 150 µm. Furthermore, their density was lower due to the presence of a greater quantity of interstitial water, which in turn contains, high concentrations of nutrients, especially ammonium [53].
In the case of the Venice lagoon, of which we have numerous data collected in different periods, significant temporal variations due to different anthropogenic impacts were also monitored.
In general, the highest sediment concentrations of N tot , P tot , P inorg and P org were recorded in Pialassa della Baiona, characterized by the highest percentage of fines (fraction < 63 µm: 87.5%) and the lowest density (0.59 g dwt cm −3 ). The same low sediment density was also recorded in the Comacchio valleys. However, the Comacchio valleys showed a significantly lower percentage of fines (57.0%), which could explain the lower concentrations of nitrogen and phosphorus in the sediments of this basin. Moreover, the poor water renewal of the Comacchio valleys accounted for the bad ecological conditions of these large ponds recorded since the last two decades of the 20th century. Despite the fact that for this basin, previous data on nutrient concentrations are missing, the list of macrophytes reported by [54] and [55] highlights that before the 1970s, the Comacchio valleys were colonized by aquatic angiosperms such as Ruppia cirrhosa (Petagna) Grande and Zostera noltei Hornemann, and sensitive macroalgae. First of all, these included Lamprothamnium papulosum (Wallroth) J. Grove, a Chlorophycea even more sensitive than aquatic angiosperms and currently present only in some fishing valleys of the Venetian lagoon such as the Cavallino valley, which shows almost pristine ecological conditions [43]. The presence of that vegetation was linked to a low trophy and nutrient concentrations certainly lower than the current ones. Indeed, in the following years, almost all the macrophytes disappeared from the valleys and only cyanobacteria and phytoplankton were recorded [44][45][46]. Discharges from fish farms, especially from eel breeding, the polluted water drainage from extensive agricultural practices and freshwater inputs from the Reno river have irremediably changed the trophic status of these basins and eliminated the macrophytes present, even those of low ecological value. Recently, some small thalli of some tionitrophilic Chlorophyceae, mostly Ulvaceae, have been recorded at only 5-15 cm from the surface because the waters were very turbid and the transparency never exceeded 25-40 cm all year round.
On the whole, the Marano-Grado and Venice lagoons were the less eutrophicated basins, showing the lowest mean concentrations of P tot (334 and 403 µg g −1 , respectively) and N tot (1.23 and 1.19 mg g −1 , respectively). The Comacchio valleys had intermediate concentrations.
However, for the Venice lagoon, many data are available from the mid-20th century ( Table 5). The concentrations of P tot and N tot recorded in 1948-49, before the development of the industrial area of Porto Marghera, were very low and can be considered the background values for the lagoon. Indeed, the mean value of the P tot concentration was only 24 µg g −1 , whereas N tot was 1.00 mg g −1 [49]. Twenty years later, both nutrients were recorded by the same authors at the same stations to have increased markedly. Total phosphorus increased ca. 6.8 times (164 µg g −1 ) and N tot almost doubled (1.86 mg g −1 [49]). The highest P tot concentrations were recorded in 1983, with a mean value of 454 µg g −1 [50], ca. 19 times greater than the value recorded in 1948-49, whereas N tot reached the highest value in 1968-73, with a mean value of 1.86 mg g −1 [49]. However, single stations close to the industrial area of Porto Marghera reached 1102 µg g −1 for P tot in 1987-88 and 4.9 mg g −1 for N tot in 2003. Indeed, after the Second World War, the lagoon received sewage effluents from about 300,000 inhabitants of Venice, Mestre, Marghera and Chioggia, wastes from chemical and steel plants located in the 30-km 2 Porto Marghera industrial area and runoff from an intensively-cultivated drainage basin (ca. 2000 km 2 ). Particularly relevant was the massive extraction of phosphorus from phosphorites and the industrial synthesis of ammonia to support the rapidly expanding agricultural production [56,57]. The main consequence was the high pollution of water and sediments. Phosphorus rapidly accumulated in surface sediments until the end of production in the mid-1980s. In the following years, P tot showed a reverse trend but its concentrations fluctuated between 403 and 412 µg g −1 because this element remained trapped in the sediments due to its sedimentary cycle. In contrast, N tot changed slightly. It was higher in 2003 (1.42 ± 1.42 mg g −1 ) and decreased markedly in 2011 (1.19 ± 1.95 mg g −1 ) during the maximum clam fishing activities [2] that favored nitrogen release in the water column as ammonium and loss into the atmosphere due to the high denitrification processes [58]. In the following years, when clam fishing activities were negligible, N tot increase again.
The studies carried out in the central part of the Venice lagoon are more recent but they show the variations of nutrients in the sediments (Table 2, Figure 4) during the various scenarios that have characterized the lagoon from the 1980s to today. In addition, for the central lagoon, P inorg and P org were also determined, highlighting the significant changes of P org while P inorg remained almost unchanged. The concentrations of P org and N tot were mainly affected by the presence of a high biomasses of macroalgae and the intensity of clam-fishing activities, which had an inverse impact on nutrient concentrations. Macroalgae produced great quantities of decomposing organic matter, which enriched surface sediments with both N tot and P org , whereas biomass reduction and fishing activities occurring between 1995 and 2012, plowing the sediments to a depth of 15-20 cm, which released into the water column the ammonium and reactive phosphorus present in interstitial waters. As a consequence, the N tot (−73%) in 2011 and P org (−51%) in 2003 reached their minimum concentrations (Figure 4). In the following years, when clam fishing was negligible or completely absent, both N tot and P org increased again but without reaching the concentrations recorded in 1987 due to the negligible macroalgal biomass present in this period. A one-way ANOVA confirmed that the most significant decrease of N tot was due to clam-harvesting that triggered the release of ammonium from surface sediments into the water column. In contrast, the strong reduction of the macroalgal biomass [36] was the main cause of P org decreasing. The analysis of the global data (Spearman's coefficients and PCA) confirms that nutrient concentrations were positively correlated with the sediment grain size (fines) and inversely with sediment density as recorded from many other analyses carried out in these environments [24,59].
Finally, plotting all the stations in the transposed PCA matrix highlights that the concentrations of nutrients in the surface sediments of the Venice lagoon cover all the concentrations recorded in the Po Delta, Marano-Grado, Comacchio and Pialassa della Baiona, which present lower environmental variability, as recorded in [24].
If we consider the concentrations of nutrients in the sediments of other lagoons recorded in the literature, Sørensen et al. [60] reported the concentrations of N tot and P tot in the Keta lagoon (Ghana) and compared their results with those of other lagoons around the world. In the Keta lagoon, P tot and N tot were in the ranges of 130-240 µg g −1 and 0.1-1.0 mg g −1 , respectively. These values were similar to the values recorded in a lagoon of the Balearic islands, Spain [61] and in the Ringkjøbing Fjord, Denmark [62] ( Table S2, Supplementary Material), but were significantly lower than the values recorded in this study.
Nutrient concentrations similar to those recorded in the TWS of the northwestern Adriatic Sea were recorded in a system of eutrophic lagoons in northwestern Mexico [63], where P tot and N tot were in the ranges of 341-1240 µg g −1 and 0.15-8.89 mg g −1 , respectively. High nutrient values were also recorded in the Méjean-Pérols lagoon (France) [64], with P tot and N tot concentrations in the ranges of 398-1147 µg g −1 and 1.9-6.0 mg g −1 , respectively. Intermediate concentrations were found in some lagoons of the Sabana-Camagüey archipelago at Cuba [65]. In these basins, P tot and N tot showed concentrations up to 241 µg g −1 and 11.5 mg g −1 , respectively. Therefore, the TWSs of the northwestern Adriatic Sea are placed among the most eutrophic basins and have concentrations strongly linked to the grain-size of the sediment. The Venice lagoon, due to its large surface and the heterogeneity of its micro-environments, presents conditions that cover the concentrations found in most of the other basins.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/w13202914/s1, Table S1. Mean, Std, Min, Max values of chemico-physical parameters in the Italian studied la-goons. Table S2. Average concentrations of P tot and N tot in the Italian lagoons and in other world transition systems.