Differences in sediment nitrogen and phosphorus distribution between riverside and lakeside wetlands in a river-connected lake

The distribution of nutrients in sediments is the result of multiple factors, including hydrological conditions and vegetation regulation, and in wetlands with complex hydrological conditions, this distribution is uncertain. In this study, the spatial distribution patterns of nitrogen and phosphorus in sediments were studied in the riverside and lakeside wetlands of Dongting Lake, a typical river-connected lake. The results showed that the nitrogen and phosphorus concentrations in the surface sediments were higher than those in the subsurface sediments in both the riverside and lakeside wetlands. In addition, the concentration of total nitrogen (TN) of lakeside wetlands in the surface sediments was higher than that of riverside wetlands, whereas the concentration of total phosphorus (TP) did not differ between the two wetland types. In the surface sediments, there were significant positive correlations between the TN and TP concentrations in the riverside wetlands (p < 0.05), but no significant correlation in the lakeside wetlands (p > 0.05). In riverside wetlands, the nitrogen and phosphorus concentrations showed strong spatial dependence. However, in the lakeside wetlands, the spatial dependence of NO3 −-N was strong, that of NH4 +-N was moderate, and that of TN and TP was weak. This study shows that both hydrology and vegetation cause differences in the distribution of nutrients in the sediments. The results obtained from this investigation clarify the differences of sediment nitrogen and phosphorus distributions in the two types of wetlands and provide a technical reference for the management of different wetland types.


Introduction
Wetlands are known as the 'kidneys of the earth' because of their ability to collect nutrients via soil erosion in the watershed and degrade nutrients via water-sediment-organism interactions [1,2].The sediment accretion rate in Florida Everglades exceeds 1 cm yr −1 [3].The net accretion rate of sediment in Dongting Lake during 2018-2019 was 3.15% [4].Rich nutrients, including nitrogen and phosphorus, accumulate in the sediments.Under the action of water flow, nutrients in sediments may be released into the water, causing eutrophication [5,6].For example, surficial sediments in Lake Winnipeg have actively and persistently released nutrients into water for many years [7].Shayo and Limbu (2018) found that sediments can be a significant source of nutrients gathering in the overlying waters, thereby intensifying water eutrophication [8].Therefore, sediment nutrients are important influencing factors of nutrient levels in water bodies.
Hydrology is the primary driving force behind changes in wetland ecosystems [9].On the one hand, hydrological conditions can directly affect nutrient distribution in sediment.Hydrodynamics change the internal nutrient release in sediments and influence the existing sediment nutrient distribution [10,11].Hydrology alters nutrient deposition in water and affects nutrient budgets in sediments, owing to the interaction of nutrients between water and sediment [12].Compared to sediment under medium hydrological connectivity conditions, the concentrations of soluble reactive phosphorus and nitrate decreased significantly in sediments with high hydrological connectivity [13].
Furthermore, hydrological conditions can indirectly affect nutrient distribution by regulating plant growth and distribution [14,15].Each plant has different requirements for hydrological conditions such as inundation frequency, duration, and water depth [16,17].Compared to Phragmites australis, Triarrhena lutarioriparia favors locations with shorter waterlogging times and shallower water depths [18].Vegetated areas exhibit different sediment nutrient distributions because of their different absorption characteristics [19,20].For example, the total accumulated amounts of nitrogen and phosphorus in Zizania latifolia were up to 135 and 18.5 grams per square meter, respectively, while those in Juncus effusus were below 30 and 5 grams per square meter, respectively [21].The nutrients absorbed by plants are returned to the sediments in the form of litter [22].One study found that litter decomposition was dependent on plant species, wetland type, and hydrological conditions, and the breakdown of Nyssa aquatica was faster than that of Nyssa sylvatica in the same wetland area [23].Therefore, the distribution of nutrients in sediment is a result of the joint action of hydrology and vegetation.Hydrology directly or indirectly dominates the input and output of nutrients and affects the distribution pattern of nutrients in the sediments.However, the distribution of sediment nutrients under different hydrological conditions is unclear, particularly in river-connected lakes.
Dongting Lake is a representative river-connected lake, which is not only connected with the Yangtze River but also with the four rivers (Xiang, Zi, Yuan, and Li Rivers) of Hunan Province.Under long-term hydrological effects, there are two wetland types in the lake: the riverside near estuaries and the lakeside far from estuaries.Owing to the great differences between the hydrological conditions of riverside and lakeside wetlands, sediment nutrients should exhibit their own distribution patterns.Therefore, an investigation of the sediments in riverside and lakeside wetlands is still needed to identify the differences of distribution patterns of nitrogen and phosphorus.The primary objectives of the this study were (1) to investigate the spatial distribution of nitrogen and phosphorus in sediments in the littoral zone of Dongting Lake, and (2) to compare and analyze the distribution characteristics of nitrogen and phosphorus in sediment in riverside and lakeside wetlands.

Study area
Covering an area of approximately 2,625 km 2 (28°38′-29°45′N, 111°40′-113°10′E), Dongting Lake is the second largest freshwater lake in China, and the Yangtze, Xiang, Zi, Yuan, and Li Rivers supply it with water.The Yangtze River connects to Dongting Lake through three inlets, Songzi, Taiping, and Ouchi, and Dongting Lake discharges water into the Yangtze River through one outlet, Chenglingji.The lake is geographically divided into three parts, namely East, South, and West Dongting Lakes, with East Dongting Lake being the largest in area.The lake has a typical subtropical monsoon climate, with an average annual precipitation of 1,200-1,415 mm, and the flooding season lasts from May to September [18].

Field investigation and sample processing
In April 2021, two riverside wetlands [Xinqiang River (XQR) and Ouchi River (OCR)] and three lakeside wetlands [Liumenzha (LMZ), Junshan (JS), and Rongjiawan (RJW)] were selected from Dongting Lake (figure 1).Two transects were selected in the riverside wetlands, from the estuary to the lake center, on both sides of the river.In each riverside transect, sampling points at a horizontal distance of 200 m were established, and six sampling points were selected on each side of the XQR and five sampling points on each side of the OCR (figure 2(a)).In the lakeside wetlands, a transect from the lake levee to the lake center was selected.In each lakeside transect, sampling points with a horizontal distance of 100 m were set up; six sampling points were selected in the LMZ and RJW, and five sampling points in the JS (figure 2

(b)).
A handheld BDS (UniStrong, China) was used to record the geographical information for each sampling point.Sediments were sampled using a Luoyang shovel and then vertically divided into five layers (0-20, 20-40, 40-60, 60-80, and 80-100 cm).These samples were put into ziplock bags and taken to the laboratory, air-dried in the shade, and debris such as stones, weeds, and animal residues were eliminated in the laboratory.After crushing the samples using a wooden hammer, they were sieved through a 100-mesh screen for property analysis.

Data analysis
For the riverside wetlands, the concentrations of nutrients in the sediments were presented using the average value calculated from two transects on both sides of the river.For the lakeside wetlands, the nutrient concentrations in the sediments were presented using values measured in a transect.The differences in NH 4

+
-N, NO 3 − -N, TN and TP concentrations in the sediments among the five locations were analyzed using a one-way ANOVA (p < 0.05).Whether the concentrations of NH 4 + -N, NO 3 − -N, TN and TP in the surface sediments at  0-20 cm and subsurface sediments at 20-100 cm (p < 0.05) were significantly correlated was tested using Pearson's correlation.Levene's test was applied to check the homogeneity of variances before analysis, and a log 10 -transformation was performed on the data whenever necessary to decrease the heterogeneity of variances.Semi-variation analyses were conducted to quantify the spatial heterogeneity of the nitrogen and phosphorus concentrations in the sediments.The semi-variance g (h) was calculated using equation (1).
where h is the lag distance, n(h) is the number of sampling points at distance h, and z(x i ) and z(x i + h) are the nitrogen or phosphorus concentrations at positions x i and x i + h.The Kolmogorov-Smirnov test was used to check the normality of the data.Semi-variance plots of nitrogen and phosphorus concentrations of sediments in the riverside and lakeside wetlands were fitted to Gaussian, linear, or spherical models.The parameters of R 2 values and RSS (the residual sums of squares) were used to evaluate the degree of model fit, where a higher R 2 value and lower RSS indicated a more optimized model.For the selected best-fit models, the nugget value (C 0 ) showed random variance among samples that included field variation and experimental error, the structural variance (C) showed variation attributed to spatial heterogeneity, and the relative structural variance [C/(C 0 + C)] showed the proportion of the total variance that was spatially structured.The larger the relative structural variance, the higher the spatial heterogeneity.In other words, < 25% suggests weak spatial heterogeneity, 25%-75% suggests moderate spatial heterogeneity, and > 75% indicates strong spatial heterogeneity [24].

Concentrations of sediment nitrogen and phosphorus
The NH 4 + -N concentration range in the five wetland locations was 0.05-40.56mg kg −1 , and the average was 6.40 mg kg −1 .The concentration range of NO 3

−
-N was 0.12-25.61mg kg −1 , with an average of 3.48 mg kg −1 .The concentration of TN varied within the range of 151.84-3042.15mg kg −1 , with an average of 821.23 mg kg −1 .The concentration of TP varied within the range of 35.30-914.50mg kg −1 , averaging 468.14 mg kg −1 .The average concentrations of nitrogen and phosphorus in all forms in the surface sediment (0-20 cm) were higher than those in the subsurface sediment (20-100 cm) (table 1).At most sampling points in both the riverside and lakeside wetlands, the concentrations of NO 3 − -N, NH 4 + -N, TN, and TP were the highest in the surface sediment (0-20 cm) (figure 3).
The surface sediment concentration of TN in the lakeside wetlands was significantly higher than that in the riverside wetlands, whereas there were no significant differences in the NH 4 + -N, NO 3

−
-N, and TP concentrations between the two wetland types.However, the concentrations of both TN and TP in the subsurface sediments at 20-100 cm in the lakeside wetlands were significantly higher than those in the riverside wetlands (table 1).

Relationship between sediment nitrogen and phosphorus
In the 0-20 cm surface sediments, TN and TP in the riverside wetlands showed a significant positive correlation (p < 0.05) (figures 4(a), (b)), whereas NH 4 + -N, NO 3 − -N, TN, and TP in the lakeside wetlands were not significantly correlated (p > 0.05) (figures 4(c), (d), (e)).In the subsurface sediments at 20-100 cm, TN and TP were not correlated in either the riverside or the lakeside wetlands (figure 5).

Spatial heterogeneity of soil nitrogen and phosphorus
The nugget values (C 0 ) of nitrogen and phosphorus concentrations in the riverside and lakeside sediments were all greater than 0 (table 2).The relative structure variances [C/(C 0 + C)] of NH 4 + -N, NO 3

−
-N, TN, and TP in the riverside sediments ranged from 80.6% to 98.1%, whereas those in the lakeside wetlands ranged from 4.8% to 77.0%.In the riverside wetlands, semi-variograms for NH 4 + -N, NO 3 − -N, TN, and TP were best described by the Gaussian and spherical models and revealed strong spatial dependence.However, in the lakeside wetlands, the spatial dependence of NO 3

−
-N was strong, that of NH 4 + -N was moderate, and that of TN and TP was weak; these were best described by the Gaussian and linear models.

Distribution of sediment nitrogen and phosphorus
In this study, the average TN concentration in the surface sediment (0-20 cm) was 1.05 ± 0.66 g kg −1 , and the average TP concentration in the surface sediment was 0.49 ± 0.15 g kg −1 , which were lower than the concentrations (TN, 1.73 ± 0.79 g kg −1 ; TP, 0.89 ± 0.12 g kg −1 ) in previous studies conducted in the littoral zone of East Dongting Lake [25].This may be because the time of sediment sampling was spring in our study and winter in a previous study.In Dongting Lake, a large amount of nutrients was deposited in the sediments during the six months of flooding, resulting in higher nitrogen and phosphorus concentrations in the sediments after the flood receded [26,27].The aboveground parts of numerous plants wither in winter, supplementing a large amount of nutrients in the sediment and increasing the sediment nutrient concentration [28].Additionally, higher nutrient bioaccumulation in plants in spring than in winter contributes to a lower nutrient concentration  in sediment in spring than in winter [29].However, the average TN concentration (1.05 ± 0.66 g kg −1 ) in the surface littoral sediment was higher than the concentration (0.73 ± 0.14 g kg −1 ) in previous research carried out in limnetic sediments of Dongting Lake [30].The possible underlying reasons are that littoral sediments retain more nitrogen from non-point source pollution and receive more wetland plant litter than limnetic sediments.The highest concentrations of nitrogen and phosphorus were mainly distributed in the surface sediments of both wetland types, which is consistent with the results of most previous studies.For example, Yang et al (2007) studied the vertical distribution of nutrients in aquatic-terrestrial ecotones and found that sediment nitrogen and phosphorus concentrations significantly decreased with increasing sediment depth [31].A study of two typical marsh wetlands showed that TN concentrations were highest in the surface sediment [32].Wu et al (2022) studied the elemental stoichiometry of sediments in a wetland critical zone and found that nitrogen and phosphorus concentrations increased from the deep layers to the surface [33].The high nitrogen and phosphorus concentrations in surface sediments may be related to plant litter decomposition, microbial activity, and sediment deposition.The decomposition of plant litter brings nitrogen and phosphorus into the surface sediment, resulting in higher nitrogen and phosphorus concentrations in this layer [34].High microbial activity in surface sediments is conducive to the aggregation of nutrients in the surface sediments, resulting in high concentrations of nitrogen and phosphorus [35].In wetlands, sedimentation promotes the accumulation of phosphorus, leading to an increase in phosphorus concentrations in surface sediments [36].
For the surface sediments (0-20 cm), this study showed that the TN concentration of lakeside wetlands was higher than that of riverside wetlands, whereas the TP concentration did not differ between the two types of wetlands.Most studies have shown that wetlands act as nitrogen sinks during submergence [37][38][39].Lakeside wetlands are subject to longer inundation time and stronger nitrogen sink than riverside wetlands, resulting in higher sediment nitrogen concentration.This phenomenon may also be attributed to the dominant plants in lakeside wetlands being more luxuriant than those in riverside wetlands, with larger biomass and more litter released nitrogen into sediments [25,40].Meanwhile, in the surface sediments, the correlations between TN and TP on the riverside were positive, and there was no correlation between TN and TP on the lakeside of Dongting Lake.One of the reasons was that the sources of nitrogen and phosphorus in the riverside wetlands were relatively simple and affected by a single exogenous pollutant, whereas the sources of nitrogen and phosphorus in the lakeside wetlands were relatively complex and affected by multiple exogenous pollutants.For example, in terms of non-point source pollution, the nutrients in the riverside sediment of the Xinqiang River originate only from the Xinqiang River watershed.However, the nutrients in the lakeside sediment of Junshan may have originated from the watershed of the Yangtze, Xiang, Miluo, and Xinqiang Rivers under hydrodynamic action.Wu et al (2018) showed that the concentration of nutrients in the Xinqiang and Ouchi Rivers was lower than that in East Dongting Lake [41], which confirmed the conclusions of this study.Another reason for the different correlations between TN and TP in the two types of wetlands may be the different transport patterns of nitrogen and phosphorus.Nitrogen is predominantly in dissolved form and migrates easier with water flow, whereas particulate phosphorus tends to settle easily where the water flow is low [42].Owing to the large flow rate in riverside wetlands, both nitrogen and phosphorus are not easy to settle; however, phosphorous settles relatively easily in lower flow rates in lakeside wetlands.Therefore, riverside wetlands with higher flow rates may exhibit a stronger correlation between TN and TP.
For sediments, the vertical distribution of nutrients reconstructs the process of nutrient input at the watershed scale [43].The concentrations of TN and TP in the surface sediments of both the lakeside and riverside wetlands were higher than those in the subsurface sediments, indicating that pollution was relatively severe in recent decades.Wang et al (2021) identified a continuous high nitrogen input and deposition in West Dongting Lake since the 1970s [44].The nutrient-loading increased fast in Dongting Lake over the past decades as a result of rapid socioeconomic development of the surrounding watershed [45].These were consistent with the conclusion of this study.

Spatial heterogeneity of sediment nitrogen and phosphorus
According to our study, except for TP in the lakeside wetlands, the nugget values (C 0 ) of sediment nitrogen and phosphorus were positive and low, indicating that there were few positive base effects caused by sampling error, short-distance variation, random and inherent variation, and that natural factors affecting sediment nitrogen and phosphorus distribution may play a principal role.The spatial heterogeneity of nitrogen and phosphorus was greater in the riverside wetlands than in the lakeside wetlands, and one reason for this was that the hydrological conditions in the riversides were relatively complex, and the water flow velocity was fast, whereas the hydrological conditions in the lakesides were relatively consistent, and the water flow velocity slow.The study shows that hydrology was one of the main factors affecting the change of nitrogen and phosphorus in sediments, and the hydrology of riverside wetlands was complex and heterogeneous [46,47].The high hydrological heterogeneity of rivers leads to a high nutrient heterogeneity in the riverside wetland sediments.
The other reason was that varied vegetation contributed to the differences in the distribution of sediment nutrients.The field investigation in this study revealed that the dominant vegetation in lakeside wetlands are Phragmites australis, Miscanthus lutarioriparius and Carex brevicuspis, and the dominant vegetation in riverside wetland are Phalaris arundinacea and Artemisia selengensis.The studies have showed that vegetation significantly affects the selection, absorption, and fixation of nutrient elements, which have significant impact on sediment nitrogen and phosphorus concentrations [48,49].Therefore, the differences in both were mainly affected by natural factors such as hydrology and vegetation.

Conclusions
In this study, the distribution characteristics of sediment nitrogen and phosphorus in the riverside and lakeside wetlands were compared by determining the concentrations of NH 4 + -N, NO 3 − -N, TN, and TP in 187 samples collected from East Dongting Lake.The results showed that the nitrogen and phosphorus concentrations in the two types of wetlands were higher in the surface sediments at 0-20 cm.The sources of nitrogen and phosphorus in lakeside sediments are relatively complex and considerably affected by exogenous pollution compared to riverside sediments.Both riverside and lakeside wetlands have been subjected to severe pollution over the past several decades.In addition, owing to the differences in hydrology and vegetation between lakeside and riverside wetlands, riverside wetland sediments had higher spatial heterogeneity of nitrogen and phosphorus than that in the lakeside wetlands.
Ammonium nitrogen (NH 4 + -N) and nitrate nitrogen (NO 3 − -N) concentrations in the sediments were detected via extraction with a potassium chloride solution (HJ 634-2012).The concentration of total nitrogen (TN) in the sediment was digested using the semi-micro Kjeldahl method (NY/T 53-1987) and determined with a continuous flow analyzer (Auto Analyzer 3 HR, Seal Analytical, Germany).The total phosphorus (TP) in the sediment was digested with HClO 4 -H 2 SO 4 and determined using molybdenum-antimony resistance colorimetry (NY/T 88-1988).

Figure 1 .
Figure 1.Study area and sampling locations.

Figure 2 .
Figure 2. Schematic of sampling point setting (circles, sampling points in the riverside wetlands; triangles, sampling points in the lakeside wetlands).

Figure 3 . 4 +
Figure 3. Spatial distributions of sediment nitrogen and phosphorus in the riverside and lakeside wetlands.The letters a, b, c, and d represent NH 4 + -N, NO 3 − -N, TN, and TP, respectively.

Figure 4 .
Figure 4. Pearson correlations of nitrogen and phosphorus concentrations in surface sediment in five wetland locations.The letters a,b, c, d, and e represent XQR, OCR, LMZ, JS, and RJW, respectively.* Significant correlation at p < 0.05; ** Significant correlation at p < 0.01; *** Significant correlation at p < 0.001.

Table 1 .
Concentrations of NH 4 + -N, NO 3 − -N, TN, and TP in the riverside and lakeside wetlands (mg kg −1 ).Different lowercase letters indicate differences between the two wetland types at a significance level of 0.05, based on Tukey's test.

Table 2 .
Best-fit models and their parameters to the semi-variogram of NH 4