Temporal and spatial distribution characteristics and source analysis of dissolved organic nitrogen in the surface sediment of the Caohai Lake, southwest China

Caohai Lake is a typical natural lake-type karst plateau wetland, which is sensitive to environmental changes, and the release of dissolved organic nitrogen (DON) in sediment is a potential factor of nitrogen pollution in this lake. Therefore, we aimed to reveal the temporal and spatial distribution characteristics and source analysis of DON in karst lake sediment. Surface sediment samples were collected from Caohai Lake in southwest China, and the sample locations were distributed across almost the entire wetland area, during both the wet and dry seasons. The DON content and fluorescence spectroscopy were determined and analysed using a three-dimensional fluorescence-parallel factor analysis (3DEEM-PARAFAC) model. The mean DON content in the sediment during the dry season (67.79 ± 42.96 mg/kg) was higher than that in the wet season (28.57 ± 20.97 mg/kg). There were four fluorescent components of DON in the sediment: C1 (tyrosine-like protein), C2 (terrestrial humus), C3 (tryptophan-like protein) and C4 (terrestrial humus). The DON in the surface sediment of Caohai Lake was influenced by terrestrial sources and biological activities. These findings help reveal the circulation mechanism of endogenous nitrogen in the lake and provide theoretical support for the prevention and control of nitrogen pollution in Caohai Lake.


Introduction
Nitrogen released from sediment is an important source of dissolved organic nitrogen (DON) in lake water (Hohman et al. 2021). When exogenous pollution is effectively controlled, endogenous release (processes such as nitrogen exchange between lake sediment and overlying water) becomes an important potential factor in water quality deterioration and the occurrence of water blooms in eutrophic lakes . DON is an important part of the nitrogen cycle in lacustrine ecosystems, and its mobility and availability play important roles in the dynamic processes of nitrogen mineralisation, fixation, leaching and plant absorption (Czerwionka 2016;Murphy et al. 2000). DON is a source of many bioavailable nitrogen components for algae, and nitrogen input increases the abundance of phytoplankton and the frequency of harmful algal blooms (Jef et al. 2018;Lin et al. 2021;Sinha et al. 2017;Sipler and Bronk 2015). In addition, DON is the main component of total dissolved nitrogen (TDN) in lake sediment (Berman and Bronk 2003), and can be converted into bioavailable nutrients through microbial action, affecting the health of lacustrine ecosystems. Generally, DON in sediment, as an intermediate nitrogen pool, plays a dynamic equilibrium role with inorganic nitrogen in the overlying water-water interface-sediment migration and transformation pathway, and can also directly participate in complex nitrogen cycles (Yan et al. 2021). Therefore, DON released into the overlying water from sediment exhibits its biological effectiveness on phytoplankton, affects aquatic ecosystems and promotes lake eutrophication to some extent (Berman and Chava 1999;Su et al. 2016).
The main compounds of DON in sediment include amino acids, urea, nucleic acids and humus (Sipler and Bronk 2015;Zheng et al. 2021), which can be characterised using UV-Vis and three-dimensional fluorescence excitation-emission matrix (3DEEM) technologies. These technologies are simple and sensitive and have become effective and widely used methods for analysing the fluorescence characteristics of DON in sediment of various aquatic environments such as rivers, lakes and reservoirs (Fu et al. 2021;Zhang et al. 2020). Parallel factor analysis (PARAFAC) is a common method for studying DON in sediment, and a three-dimensional fluorescence-parallel factor analysis model (3DEEM-PARAFAC) was established to differentiate between dissolved organic matter (DOM) in natural water by combining 3DEEM techniques (Stedmon et al. 2003). This parallel factor analysis, which can quantitatively and qualitatively analyse the 3DEEM, separates the fluorescence components into a single component, obtains the maximum fluorescence intensity corresponding to each fluorescence component and reduces the error of the instrument itself through the standardisation of fluorescence intensity. This model has attracted extensive attention from researchers and has been applied to the characterisation of DON in various natural waters and sediment (Jacquin et al. 2017;Kang et al. 2021).
Caohai Lake, in southwestern China, is a typical karst semi-enclosed lake and a natural lake-type wetland. Karst areas are more sensitive to changes in nitrogen content than non-karst areas, and many lakes in karst areas have been affected by human activities to different degrees. The results of recent studies on the sediment of Caohai Lake show that the nitrogen content of the sediment ranges from 1.94 to 14.07 g/kg, with a mean value of 7.76 g/kg. The total exchangeable nitrogen content is high, ranging from 1.70 to 7.59 g/kg, accounting for 68.7% of total nitrogen (TN), with a higher risk of potential nitrogen release (Wu et al. 2019). The total nitrogen content in the sediment has reached serious ecotoxic levels (Guo et al. 2012), and the nitrogen in the sediment of Caohai Lake during the dry period shows the release characteristics of "source" (Han et al. 2017). Therefore, it is urgent to improve the environmental conditions of Caohai Lake. However, controlling the endogenous release of sediment pollution is an exceptionally difficult task. Previous studies have focused on the forms and temporal and spatial distribution of total nitrogen in the sediment of Caohai Lake; consequently, the sources and content of DON in the sediment of Caohai Lake are poorly understood. This study adds to the information of Caohai Lake nitrogen cycle by investigating the temporal and spatial distribution of DON in the sediment of Caohai Lake and identifies the main sources that contribute to the nitrogen content of the lacustrine environment. This study can improve our understanding of the cycling mechanism of endogenous nitrogen in lakes and provide theoretical support for the management of nitrogen-derived eutrophication. The study helps to develop management strategies that are more relevant to the needs of various types of lakes and provides a reference for a more comprehensive understanding of the karst region.

Study area
Caohai Lake (26° 49′-26° 53′ N, 104° 12′-104° 18′ E) is in Weining County, Bijie City, Guizhou Province, southwest China. The upper headwater lake of Luoze River is a part of the Yangtze River system. Its watershed area is 19.8 km 2 in the dry season and 26.0 km 2 in the wet season. The watershed has a storage capacity of 3.9×10 7 m 3 (Cao et al. 2021), of which the maximum water depth is 5.0 m and the average water depth is 2.4 m (Xia et al. 2020). The wet season is from May to October, during which 87.3% of annual rainfall occurs, whereas the dry season is from December to April (Zhang 2012). The main source of lake water is precipitation, followed by groundwater (Cao et al. 2016). The lake hosts a diverse array of aquatic plants, whose coverage rate exceeds 80% (Wu et al. 2020). The emergent plants are primarily distributed in the eastern region, while the other regions are primarily dominated by submerged plants. Caohai Lake is also an important wintering place and migration centre for birds, and is known as "The Kingdom of Birds" (Peng et al. 2018). Four rivers enter, and one river exits the lake (Fig. 1). The population of the Caohai Town is concentrated in the north east area of Caohai Lake, with a few villages distributed in other areas along the lake. Disturbances by agricultural and livestock activities have resulted in extensive non-point source input of nitrogen in Caohai Lake ).

Sample collection and pre-processing
Forty-eight sampling points were selected in Caohai Lake area using the grid distribution method (Fig. 1), whereas the actual number of samples in each period varied owing to the difference in water level conditions. Surface sediment samples (0-10 cm) were collected in July 2019 and December 2019. A total of 42 and 39 samples were collected during the wet and dry season, respectively. All samples were placed in Ziplock bags, refrigerated at −4°C and returned to the laboratory. The samples were freeze dried, ground, sieved over 100 meshes, and then stored in brown polyethylene bottles at room temperature (25°C).
The sediment samples were extracted with 1 mol/L KCl solution (water:sediment = 10:1) and oscillated in a constant temperature oscillator for 1 h (condition: 25 ± 1°C, 200 rpm), and then centrifuged (5000 rpm for 10 min), supernatant was filtered through a 0.45-μm glass fibre filter membrane to remove impurities from the supernatant Shi 2018). The glass filter membrane was previously roasted at 450°C for 3 h in advance in a muffle furnace to remove impurities in the filter. The extraction solution was stored at less than 5°C in cold storage (the maximum storage time was 7 days).

DON analysis
The KCl extraction method was used to extract nitrogen from the sediment, and the DON content was calculated using the subtraction method (Eq. 1).
where ω(TDN) is soluble total nitrogen content was measured using alkaline potassium persulphate oxidationultraviolet spectrophotometry. ω(NH 4 + -N) is ammonia nitrogen content and was measured using sodium reagent spectrophotometry. ω(NO 3 − -N) is nitrate nitrogen and was (1) measured using phenol-disulphonic acid spectrophotometry. ω(NO 2 − -N) is nitroso nitrogen and was not determined, as the NO 2 − -N content in the surface sediment of Caohai Lake was nearly zero.
To verify the reliability of the experiments, 10% of the total samples were taken and the experiments were repeated.

UV-vis spectrum analysis
The scanning wavelength range was 239 to 800 nm, and the scanning step was 3 nm. SUVA 254 and E 3 /E 4 were calculated from the results of the absorbance measurements, where SUVA 254 indicates the aromaticity and degree of aromatic polymerisation of DON, and E 3 /E 4 indicates the concentration of fulvic acid and humic acid in DON.

Three-dimensional fluorescence measurement
The excitation wavelength (E x ) range was 239-800 nm, the emission wavelength (E m ) was 247-800 nm, and the scanning step Fig. 1 Geographical map of the study area and sampling points length was 3 nm and 1.17 nm. After pre-treatment, the samples were measured using a computer, and the obtained data were semi-quantitatively analysed using the PARAFAC model.
The 3DEEM fluorescence matrix group of the samples was processed by PARAFAC in the Matlab2018a DOMFluor toolbox, and the 3DEEM-PARAFAC model was established for quantitative and qualitative analysis of the fluorescence spectra. PARA-FAC is a mathematical model based on the trilinear decomposition theory combined with the least squares method. The three-dimensional data matrix X was decomposed into three matrices: A (score matrix), B (load matrix) and C, given by the model Eq. 2: When the PARAFAC model parses the 3DEEM spectrum, X ijk in the equation refers to the fluorescence intensity of the ith sample at the jth E x of k. ε ijk is the residual matrix, indicating the unknown signal; F indicates the effective fluorescence fraction; a if , b jf , and c kf denote the fluorescence component content, E x , and E m in the load matrix and are denoted as A, B, and C, respectively.
The E x /E m data from the results of running the 3DEEM-PARAFAC model were used to calculate the 3DEEM index: fluorescence index (FI), humification index (HIX) and biological index (BIX). FI can characterise the source of fulvic-like acid fluorescent substances, HIX can evaluate the humification degree of DON fluorescent substances and BIX can evaluate the relative biological contribution rate of DON fluorescent substances.

Data analysis
UV-Vis spectra, 3DEEM spectra and related parameters were measured using an AquaLog UV-800-C (HORIBA Scientific, Japan) three-dimensional fluorescence spectrometer. The 3DEEM data were processed using PARAFAC in the Matlab2018a DOMFluor toolbox. Figures were drawn using origin2018 (Origin Lab, Northampton, MA, USA) and Arc-GIS10.6 (Esri; Redlands, CA, USA), and IBM SPSS Statistics 25 (IBM, Armonk, NY, USA) was used for correlation analysis.

Temporal and spatial distribution characteristics of DON in the surface sediment of Caohai Lake
Variation characteristics of the DON content TDN content in surface sediment during the wet season ranged from 50.30 to 339.50 mg/kg, with a mean value of 189.05 ± 69.53 mg/kg. The DON content ranged from 0.77 to 91.17 mg/kg, with a mean value of 28.57 ± 20.97 mg/kg, accounting for 11.29 ± 10% of TDN. The TDN (2)  Fig. 2, the DON content in the surface sediment of Caohai Lake showed considerable seasonal variation, and the DON content in the dry season was significantly higher than that in the wet season. The correlation between DON and TDN was 0.641** in the wet season and 0.635** in the dry season, both showing a significant positive correlation (p < 0.01). In summary, it can be shown that the DON content increases with the increase of TDN, and the surface sediments of Caohai Lake are more affected by nitrogen pollution during the dry season, which was consistent with previous research (Zhou et al. 2016). In the wet season, aquatic plants were growing, and the plant rhizomes absorbed most of the nutrients, such as nitrogen and phosphorus, in water and surface sediment; in the dry season, aquatic plants began to decay with the change of climate, deposited in the form of humus and imported a large number of autogenic humus to the surface sediments. Finally, the DON content in the surface sediment in the dry season was higher than that in the wet season.

Temporal and spatial variation characteristics distribution of DON content
Areas with high DON content varied between the two seasons (Fig. 3). The areas with a high DON content were more dispersed during the wet season than during the dry season. During the wet season, the high DON content areas were mainly concentrated in the entrance and the centre of the lake and on the east, west and north shore of Caohai Lake, whereas the high DON content areas during the dry season were mainly concentrated in the central part of the lake area to the south. The overall trend was a decreasing DON content from south to north. The entrance of the lake showed a high DON content in both seasons. In terms of spatial variation, the DON content in the surface sediment of the lake entrance and the lake centre area of Caohai Lake, in the two seasons, showed high values. At the entrance of in-flowing river, this variation was due to the input of fertiliser residue from farmland, livestock discharge and domestic sewage from the lake entrance through scouring (Zhang et al. 2014). The accumulation of nitrogen source inputs from the east, west and north, resulted in high DON content in the centre of the lake area. The DON content in the surface sediment in the southern part of the dry season was relatively high. The submerged plants were mainly distributed in the southern part of the lake, likely due to the accumulation of withering and decomposition of aquatic plants in the lake during the dry season. The discharge of sewage in the surroundings also affects the nearshore nitrogen content.

Characteristics of DON fluorescence components in Caohai Lake surface sediment
The composition and source of DON can be distinguished by comparing the changes in fluorescence peak position and intensity (Fellman et al. 2010;Zhang et al. 2010). The 3DEEM-PARAFAC model was used to identify and statistically validate the fluorescence components of DON in Caohai Lake sediment (Fig. 4)  Component C1 (270/295) reflects a fluorescence profile similar to that of tyrosine (Yao et al. 2020), indicating more tyrosine-like proteins material derived from degradation by proto-endogenous microorganisms, corresponding to low and high excitation tyrosine-like fluorescence peaks (Zhu et al. 2013). Components C2 (240/420) and C4 (260, 350/470) are terrigenous humus and exist widely in natural water. These were also found in the study of pore water of Erhai (Fu et al. 2004) and Lake Hongfeng (Fu et al. 2005) water. The component C3 (275/340) reflects fluorescence characteristics similar to free tryptophan, indicating complete proteins or less degraded amino acids. This is similar to the findings of Zhang et al. (2008) in the northern Taihu Lake area, which reflected tryptophan-like proteins from microbial degradation.
The total fluorescence intensity of DON at each sampling point in the wet season was 0.55-39.98 r.u., with an average of 7.05 r.u.; the total fluorescence intensity in the dry    Stedmon et al. 2003) season was 0.94-39.84 r.u., with an average of 8.12 r.u.. The proportion of the maximum fluorescence intensity (F max ) of the fluorescence of each component (Fig. 5)  Given that the F max proportions and spatial distributions (Fig. 6) of the four fluorescent components differed very little between the two periods, the following discussion will be conducted without distinguishing between seasons. The proportion of F max of component C1 varied greatly, and the standard deviation was also large. Figure 6 demonstrates obvious spatial heterogeneity in the spatial distribution of component C1. The endogenous amino acids, represented by C1 and C3, were mainly distributed in area A and C, both of which are degraded by protein components. Previous studies showed that lacustrine algae and submerged plants are the direct sources of TN in the surface sediment of Caohai Lake, and they were the main sources of amino acids with authigenic characteristics in the DON (Dai et al. 2022;Yang et al. 2016), while the submerged plants of Caohai Lake were mainly distributed in area A and C, just as the spatial distribution of tryptophan-like and tyrosine-like represented by C1 and C3. The terrestrial sources of humus, represented by C2 and C4, were distributed throughout the lake area, mainly concentrated in areas C and D. These two areas have rivers flowing into the lake and are inhabited by farmers, and domestic sewage, agricultural production and livestock excreta pass through the surface runoff and underground seepage into the lake, becoming the source of humus in the DON (Yang et al. 2016;Zhang 2020).
The F max correlation analyses of the four fluorescent components are presented in Table 2. There was a significant correlation between each component, and the correlation between the C2 and C4 components was the most significant, with a correlation coefficient of 0.995**, which inferred that the sources of both were consistent, and the correlation between the other two fractions was significantly correlated. The correlation during the wet season was more significant than during the dry season, suggesting that the source structure of DON in the surface sediment of Caohai Lake was relatively stable. Externally, the pollutants entering Caohai Lake in different seasons were mainly agricultural fertilisers, livestock manure, and urine and domestic sewage. Internally, the sources for nitrogen pollution were mainly apoptosis and decomposition of aquatic plants and animals; therefore, the DON in the two seasons showed homology.

UV-Vis spectral features
SUVA 254 is the UV absorption value at 254 nm wavelength per unit concentration, which positively correlates with and reflects the aromaticity of DON (Lusk and Toor 2016). The ratio of the absorption values of specific bands can characterise the humification, molecular weight, and aromaticity of DON. E 3 /E 4 (A 300 /A 400 ) is the ratio of the ultraviolet absorption value of organic matter at the absorption wavelengths of 300 nm and 400 nm. These are negatively correlated with the degree of humification and molecular weight of DON and can distinguish the DON humus components in the lake (Abbt-Braun et al. 2004). When E 3 /E 4 was less than 3.5, the water-soluble organic matter in the lake mainly contained humic acid, whereas when it was greater than 3.5, it mainly contained fulvic acid.
The mean values of SUVA 254 and E 3 /E 4 were 1.17 ± 0.42 and 5.30 ± 0.49, respectively, during the wet season, whereas the mean values of SUVA 254 and E 3 /E 4 were 1.01 ± 0.46 and 6.27 ± 0.62, respectively, during the dry season (Table 3). The SUVA 254 values were slightly lower in the dry season than in the wet season. Conversely, E 3 /E 4 increased to different degrees in the dry season, indicating that the aromaticity of the DON component was slightly higher in the wet season than in the dry season. Because aromatic compounds have a stable Pi-conjugating system, DON molecules with low aromatic strength have a more fragile structure (Mladenov et al. 2021); therefore, the risk of DON release from disturbed sediment in the dry season was greater than that in the wet season. The results of E 3 /E 4 showed that DON in the surface sediment of Caohai Lake was dominated by fulvic acid, with a little humic acid in the wet season. Unlike in the wet season, the characteristics of DON in the dry season sediment were weaker aromaticity, smaller molecular weight, lower humification and higher fulvic acid content.

Fluorescence spectrum characteristics
The FI (E x = 370 nm, E m = 450 nm/E m = 500 nm) is used to characterise the source of DON. The FI values of terrestrial and biological sources are 1.4 and 1.9, respectively. When the FI is less than or equal to 1.4, the fulvic-like fluorescent substance is mainly caused by terrestrial input, whereas when the FI is greater than or equal to 1.9, the fulvic-like fluorescent substance is mainly derived from biological activities (McKnight et al. 2001). The HIX is an indicator of DON humification, when HIX is greater than 6, DON has strong humic characteristics and a large contribution from terrestrial sources, when HIX is less than 4, it indicates that DON humus characteristics are weak and predominantly autogenic, when HIX is between 4 and 6, it indicates that DON is strong humic characteristics and weakly autogenic (Zhang et al. 2010). The BIX is used to indicate the relative contribution of endogenous substances to DON; when BIX is greater than 1, it indicates that DON comes mainly from endogenous metabolism, whereas when it is between 0.8 and 1.0, the autogenous feature of DON is more significant, when BIX is between 0.7 and 0.8, DON exhibits moderately autogenous characteristics, and between 0.6 and 0.7, it indicates that DON is mainly imported from terrestrial sources, and its content is influenced by factors such as the water quality of rivers entering the lake and human activities (Huguet et al. 2009).
The FI value of the surface sediment ranged from 1.17 to 1.87, with an average of 1.42 ± 0.01 in both seasons (Table 4), indicating that the DON humus in Caohai Lake surface sediment was affected by both terrestrial input and biological activities, with the former being more dominant. However, the humus is only part of the DON, indicating that the DON is not affected only by terrestrial inputs. The change in FI with seasons was not obvious, indicating that the fulvic-like acid source structure of DON humus, in Caohai Lake surface, was relatively stable throughout the season. HIX was 0.50~5.90 in the wet season and 0.57~6.44 in the dry season. Ninety-five percent and 93% of the samples had an HIX value of less than 4 in both seasons, respectively, indicating that the degree of humification of DON is basically the same in the two seasons. The DON showed a low degree of humification and was mainly autogenic, which can be confirmed in combination with the E3/E4 analysis results discussed previously ("DON analysis").
The BIX value was 0.22~1.04 in the wet season, with a mean value of 0.79 ± 0.14, and 0.49~1.05 in the dry season, with a mean value of 0.85 ± 0.10. The mean values of BIX in the two seasons were similar, both around 0.8, which indicated that there were many new authigenic DON, and the DON in the surface sediment of Caohai Lake had obvious authigenic characteristics. This is similar to the HIX value results, and also confirms that the water environment management measures in Caohai Lake have achieved remarkable results in recent years, and the exogenous pollution has been well controlled. In the future, attention should be paid to the risk of endogenous release.

Conclusion
In this study, the temporal and spatial distribution characteristics as well as the source characteristics of DON in the surface sediment of Caohai Lake were revealed, and the following conclusions were drawn.
(1) The DON content in the surface sediment of Caohai Lake in the dry season (67.79 ± 42.96 mg/kg) was higher than in the wet season (28.57 ± 20.97 mg/kg), the dry season accounting for 32.53 ± 17.54% of TDN in sediment. The content changed significantly between the different seasons. In both seasons, the DON content in the surface sediment at the entrance and centre of the lake was higher than that in other areas.
(2) Four fluorescent components from DON were identified in Caohai Lake surface sediment: C1 (tyrosinelike protein substance), C2 (terrestrial humic acid-like substance), C3 (tryptophan-like proteins) and C4 (terrestrial humic acid-like substance). The temporal and spatial distribution of each component was stable. C1 and C3 were DON components with autogenic characteristics, mainly distributed in the lake area with high water level and concentrated growth of submerged plants; C2 and C4 were DON components with terrestrial characteristics, mainly concentrated in the southern lake area. They were close to villages, which is also the main source of untreated domestic sewage input to Caohai Lake. The correlation analysis of fluorescent components showed that C2 and C4 had strong homology.
(3) Analysis of the source of DON in the surface sediment of Caohai Lake revealed that the humic substance in DON was mainly fulvic acid, the molecular weight of DON was small, and it was easily disturbed and released into the water during the dry season with lower average water level. The low degree of humification of DON, with autogenic sources as the main source and supplemention by exogenous input, indicated that the exogenous factors of Caohai Lake pollution had been effectively controlled, and further research should be conducted around the potential endogenous release in the future.
Although qualitative and quantitative analyses of sediment DON components and sources using spectral indices were conducted in this study, the types and sources of DON fluorescence components were identified, but the specific sources of the substances could not be accurately analysed. It would be important and constructive to clarify the sources of contaminants in the sediment of Caohai Lake if the DON components could be distinguished and their contents determined at the physical level.