Quantification of Nutrient Fluxes from Sediments of Lake Hulun, China: Implications for Plateau Lake Management

Abstract

Dramatic changes in lake water conditions may induce sensitive feedback in sediments, such as the release of phosphate and ammonia. There is a lack of the understanding of sediment nutrient release fluxes and their responses to various environmental factors in plateau lakes. In this study, nutrient contents in the water and sediment, sediment nutrient release rates, and their responses to various influential factors in a steppe shallow plateau lake, Lake Hulun, China, were investigated. Sediment is a large store of nutrients, with bioavailable phosphorus of 555 mg kg⁻¹ in surface lake sediments. If water became anoxic, sediments will release appreciable amounts of phosphate (2.1 mg P m⁻² d⁻¹) and ammonia (40.0 mg N m⁻² d⁻¹), three times greater than those in oxic conditions. Raised temperatures greatly increased sediment phosphate release rates, while the highest temperature (23 °C) slightly decreased ammonia release rates due to strong nitrification. Assuming a whole year of internal loading, ammonia and dissolved reactive phosphorus concentrations in water column would increase by an estimated 0.22 mg N L⁻¹ and 0.01 mg P L⁻¹, respectively, which is equal to 2 and 0.5 times of their ambient concentrations. A pH increase from 8.4 to 9.4 significantly decreased internal nitrogen and phosphorus release rates by around three times. Overall, dissolved oxygen (DO) and pH had a greater impact on sediment nutrient release from lake sediments compared to temperature, and further increased the eutrophic status in the overlying water when these three parameters varied within the normal seasonal range (temperature: 1.9–23.1 °C, DO: oxic and anoxic; pH: 7.5–9.4). With the periodic low DO and frequency of drought in the future, lake management strategies that mitigate water eutrophication, such as water diversion, may need to be considered.

1. Introduction

The release of dissolved nitrogen (N) and phosphorus (P) from lake sediments to overlying water can dramatically impact water quality of lakes. This is especially true in productive shallow lakes with enormous pools of N and P in the sediment. N and P released from sediments into overlying water during anaerobic conditions exacerbate water eutrophication, and stimulate phytoplankton productivity [1]. Simultaneously, phytoplankton in the lake water is biodegraded by oxygen-consuming micro-organisms, which results in anaerobic conditions in the lake water and sediment. Anaerobic conditions at the profundal sediment–water interface are expected to enhance the sediment release of N and P into the overlying water. Ostensibly, internal nutrient loading acts as a positive feedback, because increased phytoplankton production promotes the anaerobic conditions in the sediment that in turn promote sediment nutrient release.

The release mechanisms of sediment N and P are closely interrelated to both biotic and abiotic processes, and can be influenced by various environmental factors including oxygen, pH, temperature, redox conditions, microbial activity, hydrodynamics, etc. [2,3,4]. Potential mechanisms include resuspension of nutrient-rich sediments [5]; release of ammonia and phosphate, resulting from the degradation of organic matter [6]; release of phosphate due to ion exchange with hydroxide at higher pH [7]; and co-release of phosphate during the reduction of iron (Fe) and manganese (Mn) oxides at anaerobic and low oxidation–reduction potential conditions [8]. For instance, the aerobic condition could facilitate the release of ammonia from sediments through the microbial reduction of nitrate to ammonia [3]. An increase in temperature primarily increases microbial mineralization, corresponding to an increase in oxygen consumption and a decrease in oxidation–reduction potential, which promotes the release of P from sediments [9]. Thus, quantification of sediment N and P fluxes and their impact factors is therefore important in eutrophic lakes and essential for lake restoration programs.

Nutrient release from sediments is not typically assessed in shallow and steppe eutrophic lakes in temperate climates, because internal loading may be considered relatively insignificant compared to external loading from multiple sources [10] including inflows, soil erosion, hay degradation, wet and dry deposition, etc. However, there is a growing interest in quantifying internal nutrient loading because of the anticipated effects of the changes in water environmental conditions, especially for lakes with a large historical N or P pool in the sediment. Multiple studies have assessed the linkage between warming water [11,12], anoxia [8,13], pH increase [14] and internal nutrient loading in lakes, while few studies have investigated the effects of changing these influential factors on sediment nutrient release, or compared the effects of these different factors in a specifically eutrophic steppe lake.

The focus of this study is the eutrophic Lake Hulun, the fourth largest freshwater lake in China. Lake Hulun plays an essential role in maintaining the biodiversity of the Hulun Buir grassland and the rich resources of animals and plants, and acts as an important ecological barrier in northeast China [15]. However, this lake becomes seriously eutrophic due to the impact of high evaporation, high mean external N and P loads (2359.6 t TN yr⁻¹ and 610.3 t TP yr⁻¹) and a lack of outflows in recent years [16,17]. Lake sediment has become a large pool for both total nitrogen (TN) and total phosphorus (TP) due to their high contents (around 4109 mg TN kg⁻¹ and 979 mg TP kg⁻¹ in June 2014) [17]. In addition, the water temperature of Lake Hulun has increased slightly over the past 10 years. Between 2005 and 2018, the lake water has warmed at an average rate of 0.07 °C (Figure S2), mainly as a result of global climate change. The pH of Lake Hulun has increased slightly from 2010 to 2015, and periodic low oxygen conditions also occur in water column [10]. All the environmental changes may promote sediment N and P release, and increase the risk of water eutrophication.

We evaluated the flux rates of nutrients (ammonia, nitrate and phosphate) from the profundal sediment of Lake Hulun in light of recent changes related to water temperature, pH levels and dissolved oxygen (DO). In contrast to many studies that have focused on sediment nutrient release, particularly in anoxia or warming conditions [12,13,14], our study was more comprehensive, and evaluated the effects of both warming and anoxia as well as pH increase on internal nutrient loading. This study is valuable in that few studies have experimentally examined the impact factors of sediment nutrient release in such a eutrophic steppe lake.

2. Materials and Methods

2.1. Geological Background

Lake Hulun (48°30′40″–49°20′40″ N, 117°00′10″–117°41′40″ E) is located in the Hulun Buir steppe in northeast China (Figure 1). The lake area is approximately 2000 km², and it has a mean water depth of 5–6 m and a water storage capacity of 13.8 billion m³. Crulen River and Orshun River are the two main inflow rivers; their amounts of water have decreased in recent years. Xinkai River used to be the main outflow of Lake Hulun; however, it has become an inflow in recent years, because the lake water level has been declining [18]. Due to the increase in air temperature and decrease in precipitation in the basin, the lake area dramatically shrank from 2247 to 1760 km², with the mean water depth decreasing from 7.0 m to 2.4 m from 2000 to 2012 [18]. Afterwards, the lake experienced a rapidly increasing trend in water level, area and volume from 2012 to 2015 [19]. The lake water quality was characterized as Grade V (GB3838-2002) [20], with mild eutrophication [15]. The annual mean concentrations of TN and TP, and the ratio of TN/TP and Chlorophyll a (Chl-a) were 0.2–3.5 mg L⁻¹, 0.1–0.3 mg L⁻¹, 3.5–25.2, and 3.3–10.3 μg L⁻¹, respectively, from 2006 to 2015 [10] (see Figure S1). The mean pH in the lake water ranged from 8.8 to 9.4, and the mean DO concentration was 7.12 mg L⁻¹ over the period 2008–2015 [10].

Due to shrinking of the surrounding wetlands and deterioration of the water quality (as well as the degeneration of ecological environment), the survival of waterfowl and birds is directly threatened [21]. In addition, the lake has been fished commercially since 1950, and is now experiencing a severe decline in the number of fish and species present due to the decrease in the lake’s storage capacity and also to overfishing [22]. Furthermore, the fish breeding and cropland areas are widely distributed throughout the Zhalainoer District, close to northeast of the lake, and about 400 herdsmen and large number of livestock are also living in the Lake Hulun basin [18]. The deterioration of water quality greatly damages the freshwater supply for supporting the fish culture, agriculture irrigation, and stock farming in this basin.

2.2. Sediment and Water Collection and Incubation

Water and sediment samples were collected at Shuanmazhuang station (Figure 1) on 14 August 2018, in the southwestern region of Lake Hulun (117.15° E, 45.92° N). The water body and sediment in the southwest of the lake present high TP concentrations throughout the year [20]. On the other hand, Lake Hulun has a long ice cover period of 6–7 months, lasting from November to early May of the following year. Thus, sampling at Shuanmazhuang station in August is easy to carry out, and is representative of high internal loading under warm temperatures. Surface sediment (0–10 cm) was collected using a Peterson grab sampler, and lake water was collected using a water collector. A total of 40 kg sediment and 400 L water were collected. In the laboratory, large rocks and pieces of wood were picked out of the sediment. Then, it was sieved at 1.8 mm, air-dried to constant weight, and mixed thoroughly. All samples were stored in a refrigerator at 4 °C before analysis and experimentation.

The experimental reactor comprised polyethylene tubes with the dimensions of 20 cm × 30 cm (diameter × length). Sediment was evenly added to the reactor to a depth of 5 cm. Then, the in situ water was added slowly, and poured along the wall of the reactor to minimize sediment disruption, to a depth of 20 cm.

Three duplicate reactors were incubated under three conditions, including DO, temperature and pH. An experiment for DO effects was conducted under current, oxic and anoxic conditions in the presence of light, with a similar temperature of around 22.5 °C (Table S1). Reactors in the current condition were incubated in the natural environment, while in oxic and anoxic conditions, they were controlled by the circulation of oxygen (99%) and nitrogen (99%) for 2 h after daily sampling, respectively. It was noted that oxic treatments (mean DO concentration was 8.0 mg L⁻¹) represented the original DO conditions (7.1 mg L⁻¹) in lake water [10]. An experiment for temperature effects was conducted at 2, 5, 15, and 25 °C in constant temperature incubators, representing winter, spring, autumn, and summer water temperatures in Lake Hulun. The water temperature was controlled at approximately 1.9, 7.0, 14.8, and 23.1 °C in each treatment (Table S1). An experiment to evaluate the effects of pH was conducted at around 6.5, 7.5 and 8.5 according to the weak alkaline conditions of Lake Hulun’s water. The water pH was adjusted by 1 mol L⁻¹ NaOH and 1 mol L⁻¹ HCl. All incubation experiments were carried out simultaneously and lasted for 30 days (d).

All incubation reactors were maintained at a water volume of 6.28 L by replacing the sample water with in situ lake water. The water level in these experiments was noted in order to maintain the same water quantities after sampling and supplementation.

2.3. Sediment Analyses

Sediment TP was determined using the potassium persulfate oxidation method [23]. The sediment P fraction was performed in duplicate based on the modified chemical sequential extraction method of Rydin [24]. Loosely sorbed P (LS-P), redox-sensitive P (Fe-P), aluminum-bound P (Al-P), organic P (Org-P), carbonate and apatite P (Ca-P), and Residual P (Res-P) were determined following the procedure in Table S2. Phosphorus in different fractions was determined using the molybdenum blue/ascorbic acid method [25]. Res-P was taken as the difference between TP and all other P fractions. All acids used in the analysis were of guaranteed reagent grade. The water used for dilution, reagent preparation, and washing was pure water (Milli-Q water). The recovery (%) of different P forms was measured as the ratio of the P concentration determined by extraction to that determined by ignition. The results showed that the recovery of different P forms ranged from 90.41% to 109.93%, with an average of 99.75% (n = 3), suggesting that the contents of the different P forms in sediments was satisfactory.

Before the incubation experiment, 100 mL water was sampled to measure the concentration of total phosphorus (TP), dissolved total phosphorus (DTP), dissolved reactive phosphorus (SRP), total nitrogen (TN), dissolved total nitrogen (DTN), ammonia (NH₄⁺-N), nitrate (NO₃⁻-N), and nitrite (NO₂⁻-N) using standard colorimetric analytical methods. Particulate total phosphorus (PTP) and particulate total nitrogen (PTN) concentrations were calculated by subtracting DTP and DTN from TP and TN, respectively. The dissolved organic phosphorus (DOP) concentration was calculated by subtracting SRP from DTP. The dissolved organic nitrogen concentration (DON) was obtained by subtracting NH₄⁺-N, NO₃⁻-N, and NO₂⁻-N from DTN.

Reactor monitoring included collecting a 50 mL water sample, to measure ammonia, nitrate, and SRP. Nutrient analysis was performed in three duplicates and reanalyzed if there was a discrepancy of more than 20%. In addition, the indicators of pH levels, DO concentrations, and temperatures in the overlying water were measured using YSI (Pro Quatro) during the incubation period.

Reported flux rates of nitrate, ammonia, and SRP were estimated based on the equation below [3]:

$$R=\left[V\left(C_n-C_0\right)+\sum _{j=1}^n{V}_{j-1}\left(C_{j-1}-C_a\right)\right]/(S\times t)$$

where R is the release rate (mg m⁻² d⁻¹), V is the overlying water volume in the treatments (L), $C_n$, and $C_0$, and $C_{j-1}$ is the nutrient concentration in the overlying water at the time of n, o (origin), and j−1 (mg L⁻¹), respectively. $C_a$ is nutrient concentration in the water added to the overlying water (mg L⁻¹), $V_{j-1}$ is the sampling volume at the time of j−1 (L), S is the surface area of the experimental chamber (m²), and t is the sampling time (d). The fluxes calculated this way most likely underestimate the actual in situ flux, because the concentration of overlying water increases over time, thereby reducing the concentration gradient and thus the flux [26]. According to the measured daily water temperature from the Ecological Environment Monitoring Station in Lake Hulun, the number of days for each specified temperature scenario can be counted and multiplied by the release rate, and finally summarized to obtain the annual release amount. In terms of the observations, the number of days for water temperatures of 1.9, 7.0, 14.8, and 23.1 °C is roughly 180, 30, 125, and 30 days, respectively.

2.4. Statistical Analysis

All statistical analysis in this study was performed using ArcGIS 10.7 and Origin 2021 software.

3. Results

3.1. Water and Sediment Quality

The TN and TP concentrations at the sampling site were 2.2 and 0.21 mg L⁻¹ (Figure 2), respectively, evaluated as Grade V according to the Chinese Environmental Water Quality Standard for Surface Water (GB3838-2002). The TN and TP concentration of the lake water was around two to four times that of the standard of Grade III water quality, respectively. High N and P levels are usually accompanied by water eutrophication, which may have adverse impacts on lake ecosystems. Specifically, the concentrations of PTN, DON, NH₄⁺-N, and NO₃⁻-N were 0.45, 1.65, 0.09, and 0.02 mg L⁻¹, respectively. The concentrations of PTP, DOP, and SRP were 0.08, 0.11, and 0.02 mg L⁻¹, respectively.

The sediment of Lake Hulun was rich in N relative to P (

Table 1. The content of different forms of N and P in the sediment.

Parameter	Value (mg kg−1)	Percentage (%)	Reference
TN *	2051		[32]
TN *	2170		[33]
TP	1079 ± 21	100	this study
LS-P	184 ± 8	17.3
Fe-P	69 ± 4	6.4
Al-P	302 ± 12	28.3
Org-P	306 ± 10	28.7
Ca-P	173 ± 6	16.2
Res-P	37 ± 2	3.1
Fe *	32,000		[34]

* represents the mean content of the entire lake sediment.

), which is similar to the surveyed results in 2008 [27]. The sediment N fraction was dominated by organic nitrogen (93.6%). The sediment P fraction was dominated by LS-P, Al-P and Org-P, which made up 74.2% of the TP in sediments. The fraction of total P as bioavailable P (including LS-P, Fe/Al-P) was 52.0% at Shuanmazhuang station. The Lake Hulun sediment had a relatively high Fe/P ratio of approximately 30. Fe/P ratios > 5–10 suggest that Fe hydroxides control P retention in lake sediment [2]. The mean TN and TP contents in the Lake Hulun sediments were relatively higher compared to those in Lake Chaohu [28]. The TP at Shuanmazhuang station indicated a high level of P in Lake Hulun, with a relatively large proportion of bioavailable P (52.0%). Generally, the nutrient level and forms for both N and P in the lake sediment are related to the sediment age, long-term transport, and history of water quality and algal blooms [29,30]. The depth of the Lake Hulun sediments is more than 80 cm, representing the sedimentary history of the past 4000 years [31]. The surface 10 cm sediments, settled in recent decades, are enriched with high levels of total P and bioavailable P, which is consistent with the findings of Lü et al. (2018) [31]. These results suggest that Lake Hulun sediments could be a significant source of P to the overlying water.

3.2. Effects of Oxygen on Sediment N and P Release

Ammonia showed a rapid increase in oxic, anoxic and control treatments during the first 4 d of incubation (Figure 3a). Then, ammonia decreased sharply and formed one new balance around 0.2 mg L⁻¹ in the control and oxic treatment, while in anoxic treatment, ammonia decreased first to 0.3 mg L⁻¹, and substantially increased afterwards up to 1.31 mg L⁻¹. The mean ammonia release rates in oxic, anoxic and control treatments were 14.4, 40.0, and 23.1 mg m⁻² d⁻¹, respectively, over the whole incubation period (

Table 2. Nutrient fluxes of ammonia, nitrate and SRP in experimental treatments under different DO, T and pH conditions.

Incubation Period	Treatments	Nutrient Fluxes (mg m−2 d−1)
		Ammonia	Nitrate	SRP
DO a control	Control	1.0–73.1 (23.1)	3.0–45.2 (15.5)	0.4–6.3 (2.1)
	Oxic	0.2–50.8 (14.4)	1.8–63.0 (21.5)	0.3–2.2 (0.8)
	Anoxic	4.4–164.1 (40.0)	2.6–73.2 (21.7)	0.5–5.4 (2.2)
T a control	T1	−4.0–30.3 (9.8)	0.7–21.4 (9.1)	0.1–1.8 (0.5)
	T2	5.9–39.6 (16.3)	−2.7–27.0 (8.1)	0.1–1.8 (0.6)
	T3	−3.1–34.9 (17.2)	−0.1–26.7 (10.5)	0.2–4.0 (0.8)
	T4	2.6–37.3 (17.0)	13.6–186.5 (62.9)	0.3–4.3 (1.0)
pH a control	pH1	1.9–54.1 (20.3)	5.3–23.1 (13.1)	0.3–5.5 (2.2)
	pH2	−1.7–61.3 (21.4)	7.7–95.2 (36.0)	0.5–10.3 (3.3)
	pH3	−3.3–34.0 (7.4)	11.8–129.1 (55.2)	0.3–2.1 (0.9)

^a indicates nutrient release rate range over the experimental phase. Values in the brackets represent the mean nutrient release rates.

). Anoxic release rates of ammonia (40.0 mg m⁻² d⁻¹) were higher than those observed in sediments from lakes with trophic status (>15 mg m⁻² d⁻¹) [35].

Nitrate production, which is a byproduct of biological ammonia oxidation, can account for a significant portion of the overall oxygen consumption in bottom water. The phenomenon of oxic and anoxic accumulation was captured in our experimental sediment–water incubations. The nitrate in the oxic treatment increased continuously from an initial 0.15 mg L⁻¹ to a peak value (0.8 mg L⁻¹) during the first 7 d of incubation (Figure 3b), and then decreased slowly. The nitrate in the control treatment exhibited similar pattern. By contrast, nitrate in anoxic treatment decreased to 0.4 mg L⁻¹ at 7 d after a rapid increase over the first 2 d of incubation, and then increased slowly to the magnitude of the value between control and oxic treatments. In the anoxic treatment, the rapid nitrate increases and decreases were synchronous with ammonia (Figure 3a–c), which was similar to the previous observations of sediment from Deer Lake [13]. However, the difference with former study was that nitrate did not decrease in the anoxic condition. Thus, microbial nitrification was primed in Lake Hulun sediments, even in the anoxic condition. Using the referenced sediment oxygen demand (SOD) of 4.3 g DO per g nitrate produced during nitrification [36], the measured nitrate flux in this study accounted for SOD of 90.7 mg m⁻² d⁻¹. This amount is significantly larger compared to the incubation result in Lake Tahoe [12]. Based on this calculation, microbial nitrification seems to account for a large fraction of SOD, and contributes to nitrate accumulation in Lake Hulun.

The SRP in oxic, anoxic, and control treatments showed a steady increase over the incubation period. Compared to the control treatment, the SRP in the oxic treatment showed a lower increase rate, while high rates occurred in the anoxic treatment (Figure 3d). A dramatic drop of SRP in the 4 d to 7 d was likely associated with the absorption of SRP by particulate P. The mean release rates of SRP in oxic, anoxic and control treatments were 0.8, 2.2, and 2.1 mg m⁻² d⁻¹, respectively (Table 2). These rates are similar to those observed in other mesotrophic lakes. The evidence was that P release rates from anoxic sediments typically ranged from around <2, 2–10, and >10 mg m⁻² d⁻¹ for oligotrophic, mesotrophic and eutrophic lakes, respectively [37]. An interesting phenomenon was observed: the SRP release rates in the control treatment were similar to those in the anoxic treatment, indicating that low DO concentrations in the control treatment (average value is 5.0 mg L⁻¹) could also promote sediment P release. Thus, internal P release should be concentrated in the occasionally low oxygen condition in lake water, for example, in low mean DO concentrations (4.1 mg L⁻¹) in 2010.

3.3. Effects of Temperature on Sediment N and P Release

Ammonia concentrations in the overlying water were generally higher under 14.8 °C, with a range of 0.04 to 1.7 mg L⁻¹, compared to those in 1.9 °C (0.02 to 0.8 mg L⁻¹) (Figure 4a), while under 23.1 °C, ammonia concentrations initially increased to 0.80 mg L⁻¹ at 4 d, and then increased to 1.6 mg L⁻¹ rapidly after decreasing to a minimum value of 0.2 mg L⁻¹ at 11 d. The short-term decrease in ammonia concentrations from 4 d to 11 d may be due to the strong nitrification under the highest temperature (23.1 °C). Nitrate concentrations experienced a similar pattern. Specifically, strong positive correlations between ammonia and nitrate concentrations were observed under all treatments (r² = 0.4, 0.3, 0.7, 0.4, respectively) over the incubation period. This indicated that the cycling of the two compounds was interrelated. The mean release rates of ammonia were 9.8, 16.3, 17.2 and 17.0 mg m⁻² d⁻¹, respectively, with a water temperature of 1.9, 7.0, 14.8 and 23.1 °C.

Nitrate flux rates measured in experimental treatments averaged 9.1, 8.1, 10.5 and 62.9 mg m⁻² d⁻¹ as the temperature increased (Table 2). In the highest temperature treatment (23.1 °C), nitrate in the overlying water accumulated rapidly in the first 11 incubation days (Figure 4b), indicating that nitrification was strongly enhanced at 23.1 °C during the period. In contrast, nitrate had no obvious accumulation in relatively lower temperature treatments during the same period (Figure 4b), indicating a balance between nitrification, diffusion, denitrification, etc. Afterwards, nitrate accumulated rapidly and reached a new balance similar to that in the highest temperature treatment. These observations indicated that high temperatures can stimulate strong nitrification and result in nitrate accumulation in lake water over a short period.

When other conditions were kept constant, sediment P release increased significantly with temperature increase (Figure 4d). SRP concentrations in the overlying water increased gradually as the incubation day increased, which indicated a continuous release of SRP from lake sediments. Mean values of SRP concentration during the investigation period reached a range of 0.02 to 0.1 mg L⁻¹ with 23.1 °C, compared to a range of 0.02 to 0.09 mg L⁻¹ with 1.9 °C. Correspondingly, the release flux rates of SRP in the warmest treatment (1.0 mg m⁻² d⁻¹) were two times greater than those in the lowest temperature treatment (0.5 mg m⁻² d⁻¹) (Table 2).

Extrapolating sediment fluxes under each temperature to the entire Lake Hulun, with a sediment surface of 609 km², the annual internal N and P loading increases the ammonia and phosphate concentrations in the water column by an estimated 0.22 mg N L⁻¹ and 0.01 mg P L⁻¹, which is 2 and 0.5 times that of the ambient concentrations, respectively. In addition, the internal N and P loading was evaluated to be 2992.0 t N yr⁻¹ and 144.9 t P yr⁻¹, which is 127% and 24% of the external N and P loads [16]. In Lake Taihu, the internal nutrient loading was estimated to be 2–6 times the external loading [38].

3.4. Effects of pH on Sediment N and P Release

In alkaline conditions, ammonia concentrations decreased when the pH increased from 7.5 to 9.4 (Figure 5a) (incubation days ≤ 4). Correspondingly, the nitrate concentrations increased as the pH increased (Figure 5b). Subsequently, both ammonia and nitrate concentrations had no obvious differences under different pH conditions. Ammonia concentrations in the overlying water experienced strong fluctuation over the incubation period. The mean ammonia release rates were 20.3, 21.4 and 7.4 mg m⁻² d⁻¹, respectively, with the increase in pH. In contrast, the nitrate in all pH conditions increased steadily in the first 11 d of incubation, and slightly decreased to a similar concentration of around 1.8 mg L⁻¹. Mean accumulation rates of nitrate increased from 13.1 to 55.2 mg m⁻² d⁻¹ with the increase in pH over the whole incubation period. This can primarily be attributed to the strong nitrification when the pH increased.

SRP concentrations in the overlying water increased gradually to a peak value between 0.06 and 0.1 mg L⁻¹ at the 11d incubation, then decreased slightly, and stabilized at this level afterwards (Figure 5d). Correspondingly, SRP release rates increased initially from 2.2 to 3.3 mg m⁻² d⁻¹ as the pH increased from 7.5 to 8.4, and then decreased to 0.9 mg m⁻² d⁻¹ as the pH continually increased to 9.2. As reported in the previous study, the mean annual pH was 8.8–9.4 in Lake Hulun [10]. Thus, a release rate of about 0.9–3.2 mg m⁻² d⁻¹ can generally represent the original sediment P release rate, which was consistent with the oxic P release rate (0.8 mg m⁻² d⁻¹) and the observed rate in the highest temperature treatment (1.0 mg m⁻² d⁻¹).

3.5. Overview of Nutrient Fluxes in Shallow Lakes

Decreasing algal bloom and increasing water quality in shallow lakes is usually more difficult than in deep lakes. Eutrophic shallow lakes have typically responded slowly to reduced external nutrient loading, usually because of the longevity of internal loading. An overview of internal N and P fluxes obtained from sediment incubations under various conditions in shallow lakes was summarized for lake management (see Table 3).

Clearly, our findings in this study are in the range of reported N and P fluxes from shallow lake sediments around the globe (Table 3). Qin et al. (2006) using cores from Lake Taihu for incubations under different temperatures found a significant difference between different temperatures. They reported that the annual mean fluxes of N and P from sediments were 11.7 mg N m⁻² d⁻¹ and 1.1 mg P m⁻² d⁻¹, respectively [38]. A similar work by Zhang et al. (2006) demonstrated strong spatiotemporal variation in sediment N and P release rates in Lake Taihu [39]. They reported higher values compared to Qin, which can be attributed to the different sampling sites and sampling times, especially for Lake Taihu, which has a large area and high N and P heterogeneity in its lake sediments. Work by Wang et al. (2018) showed that the release rate of sediment P was significantly higher under a high incubation temperature of 25 °C in Lake Hongfeng [40]. The internal P release rate in Western Lake Erie was significantly increased when the sediment incubation conditions changed from oxic to anoxic; the same was true of sediments from Lake Pontchartrain [41,42]. Importantly, these estimates only considered diffusive flux in a static environment, and excluded bioturbation and resuspension induced by wind. Release rates with resuspension were up to 20–30 times that of undisturbed sediment, and were mainly related to dissolution from particles due to low Fe/P ratios and high fractions of loosely sorbed P, rather than to the quantities of particles re-suspended [43]. Thus, a wide range of sediment N and P fluxes have been reported in shallow lakes; overall, it is clear that this internal nutrient source cannot be ignored.

Table 3. A summary of measured ammonia and phosphate release fluxes from sediment static incubations in shallow lakes.

Location	Incubation Temperature (°C)	Oxic P Release Rate (mg m−2 d−1)	Anoxic P Release Rate (mg m−2 d−1)	Oxic N Release Rate (mg m−2 d−1)	Anoxic N Release Rate (mg m−2 d−1)	References
Lake Taihu	5, 15, 25	1.1		11.7		[38]
Eastern Lake Taihu	8–30	2.1 ± 1.7		44.9 ± 21.9		[39]
Northern Lake Taihu (Meiliang Bay)	8–30	0.5 ± 0.5		16.2 ± 12.0		[39]
Western Lake Chaohu	15–30			13.1–32.9		[44]
Eastern Lake Chaohu	15–30			4.5–17.4		[44]
Northwestern Lake Chaohu	25	0.1–13.0		14.3–128.2		[45]
Lake Dianchi	5, 15, 25			12.7–59.7 (30.2)		[46]
Lake Dianchi (Fubao Bay)	14–16	0.9–4.9 (2.7)		22.9–163.1 (111.7)		[47]
Lake Dongting				2.0–147.0 (16.2)		[48]
Lake Dongting	12	0.04–0.3 (0.2)				[49]
Lake Hongfeng	5, 15, 25		0.4, 0.6, 0.9			[40]
Lake Nansi		0.3–2.7 (1.1)		3.1–10.3 (7.0)		[50]
Lake Hulun	23	0.3–2.2 (0.8) (DO = 8)	0.5–5.4 (2.1)	0.2–50.8 (14.4)	4.8–164.1 (40.0)	This study
Lake Hulun	2, 7, 15, 23	0.1–4.3 (0.7) (DO = 4–6.8)		−4.0–39.6 (14.4)		This study
Western Lake Erie (America)	20	0.4 ± 0.3	9.3 ± 6.5			[42]
Lake Pontchartrain (America)	25	0.4 ± 0.1	0.9 ± 0.2			[41]
Lake Rotorua (Zew Zealand)	20.8		10.6–30.7 (16.1)		75.1–484.5 (244.3)	[51]
Lower Havel (Germany)			3.5–36.0		20.0–124.0	[52]
Swarzędzkie (Poland)	2.0–20	−2.4–59.5	2.8–26.9			[53]

Table 3. A summary of measured ammonia and phosphate release fluxes from sediment static incubations in shallow lakes.

Location	Incubation Temperature (°C)	Oxic P Release Rate (mg m−2 d−1)	Anoxic P Release Rate (mg m−2 d−1)	Oxic N Release Rate (mg m−2 d−1)	Anoxic N Release Rate (mg m−2 d−1)	References
Lake Taihu	5, 15, 25	1.1		11.7		[38]
Eastern Lake Taihu	8–30	2.1 ± 1.7		44.9 ± 21.9		[39]
Northern Lake Taihu (Meiliang Bay)	8–30	0.5 ± 0.5		16.2 ± 12.0		[39]
Western Lake Chaohu	15–30			13.1–32.9		[44]
Eastern Lake Chaohu	15–30			4.5–17.4		[44]
Northwestern Lake Chaohu	25	0.1–13.0		14.3–128.2		[45]
Lake Dianchi	5, 15, 25			12.7–59.7 (30.2)		[46]
Lake Dianchi (Fubao Bay)	14–16	0.9–4.9 (2.7)		22.9–163.1 (111.7)		[47]
Lake Dongting				2.0–147.0 (16.2)		[48]
Lake Dongting	12	0.04–0.3 (0.2)				[49]
Lake Hongfeng	5, 15, 25		0.4, 0.6, 0.9			[40]
Lake Nansi		0.3–2.7 (1.1)		3.1–10.3 (7.0)		[50]
Lake Hulun	23	0.3–2.2 (0.8) (DO = 8)	0.5–5.4 (2.1)	0.2–50.8 (14.4)	4.8–164.1 (40.0)	This study
Lake Hulun	2, 7, 15, 23	0.1–4.3 (0.7) (DO = 4–6.8)		−4.0–39.6 (14.4)		This study
Western Lake Erie (America)	20	0.4 ± 0.3	9.3 ± 6.5			[42]
Lake Pontchartrain (America)	25	0.4 ± 0.1	0.9 ± 0.2			[41]
Lake Rotorua (Zew Zealand)	20.8		10.6–30.7 (16.1)		75.1–484.5 (244.3)	[51]
Lower Havel (Germany)			3.5–36.0		20.0–124.0	[52]
Swarzędzkie (Poland)	2.0–20	−2.4–59.5	2.8–26.9			[53]

4. Discussion

4.1. Mechanisms of Nutrient Release from Lake Sediments

Profundal sediment in Lake Hulun stores a large amount of N and P that will be released to the overlying water under anoxic, high-temperature and weak alkaline conditions (Table 1 and Table 2). Because several of these mechanisms may interact at the same time and vary in importance from lake to lake, there has been little success in determining their relative importance. Based on observations of laboratory experimental data (Table 2), we have also made judgements about their relative importance in Lake Hulun.

Typically, ammonia release from lake sediments in anoxic conditions is due to a loss of biological nitrification and a decrease in ammonia assimilation by anaerobic microorganisms (Figure 6) [35]. In contrast, release of P from sediments under anoxic conditions is attributed to iron reduction due to lower redox conditions in the sediment [41]. In the present study, we found that anoxic treatments could significantly stimulate sediment ammonia and phosphate releases approximately three times greater than those in the oxic environment in Lake Hulun (Table 2). Similarly, such a case was also documented in some Danish lakes, in which a short-term nutrient increase rate of over 100 mg P m⁻² d⁻¹ was observed in anoxic conditions [2].

The release of sediment N and P at higher temperatures can be explained by the fact that microbial vitality, bioturbation, mineralization, and anaerobic biological conversion were enhanced at increased temperature (Figure 6) [54]. As the air temperature increased, the biological rate increased, and a large amount of DO was consumed. The bottom water of the lake became anoxic and accumulated a certain amount of bioavailable nutrients; this likely increased the sediment nutrient release [12]. On the other hand, the organic N and P in sediments were converted through intensified microbial activity into inorganic N and P, and then released. In this study, the similar sediment N release rate between the two higher temperature treatments was due to the high nitrification rate (Table 2 and Figure 4).

High pH promoted the release of Fe and Al-bound phosphorus into the sediment (Figure 6) [14]. The P-binding capacity of Fe and Al compounds decreases as pH increases in the overlying water and sediment, primarily due to the ligand exchange reactions that cause orthophosphate to be replaced by hydroxide ions [55]. The high pH (10–11) caused by algae photosynthesis in the water column and in the upper 10 cm of sediments maintained the high SRP concentrations in Lake Søbygaard [56]. The P release from the Lake Gonghu, East Taihu, and Meiliang sediments was more favorable in alkaline conditions, as pH increased from 8 to 12 [14]. However, contrary to previous studies, the slight pH increases in Lake Hulun decreased the sediment N and P release rates by about three times (Table 2). The potential mechanism responsible was probably the absorption of phosphate by CaCO₃, or co-precipitation with Ca(OH)₂ [57]. Similarly, little phosphate was released from Lake Swan sediments when the pH was 10, while a large amount of P was released when the pH was 9 [58]. Overall, the main factors affecting sediment N and P release rates in Lake Hulun are DO and pH, according to the absolute N and P release rates from sediment incubations under various conditions within a certain range (Table 2).

4.2. Implications for Public Health and Lake Management

The relatively large internal N and P release rates have considerable implications for public health. Internal N and P release can support the growth of phytoplankton and cause significant hysteresis in ecosystem responses when the external nutrient loading is reduced [59]. The algae bloom will break out when there is excessive algal growth. The outbreak of algae bloom consumes a large amount of DO in the water body, even causing the death of aquatic organisms and an imbalance of the lake’s ecosystem [60]. In addition, the alkaline water (pH > 8) of Lake Hulun promotes the growth of phytoplankton, and especially of blue-green algae [61]. Although less toxic Microcystis is present in the Lake Hulun water, dominant species such as Cyclotella meneghiniana in Bacillariophyta may accumulate in the coastal lake zone for a long time, which may cause the water body to become anoxic, and even black and smelly [60]. Furthermore, Lake Hulun is an important habitat for birds, fish and other creatures, and a water source for livestock and residents in the grassland, the factories along the upper reaches of the inflows (Crulen River and Orshun River), and several large fishing grounds (such as Xiaohekou fishing ground) around the lake; it is also a tourist destination [10]. The high N, P and Chl-a concentrations near the inflow inlets (Crulen River and Orshun River) and fishing grounds, such as Wudulu fishing ground in the southeast of Lake Hulun, will come with great risk to the health of the lake’s ecosystem and the sustainable development of the region.

The relatively large sediment N and P release fluxes have considerable implications for lake management measures. The endogenous nutrient source is difficult to control. Some restoration measures such as Al salt addition, sediment capping, and dredging are not feasible in such a large steppe lake, due to their need for great investment, their adverse impacts on other organisms, and their significant requirements in terms of dredging methods [62]. As the sediment legacy of N and P is largely derived from nutrient loading from exogenous sources such as inflows, atmospheric deposition, hay input, runoff, and tourist industry [10], management actions to reduce the external inputs from above sources are expected to be implemented. The feasible measures include protecting the grassland in the lake basin, harvesting dead hay in a timely manner, strictly prohibiting excessive grazing, reasonably using livestock manure, reducing surface runoff with large amounts of nutrients, and strengthening the treatment and supervision of factory wastewater in the upper reach of inflows; these measures are recommended to protect the water quality of Lake Hulun.

Increased outflow discharge and water diversion from the Xinkai River can also reduce the nutrient concentration to a certain degree, through enhancing the water exchange rate and flushing out nutrients from the lake [29]. The biggest advantage of water diversion is that lake quality may show a quick response in nutrient reduction when suitable dilution water is available; however, this implementation needs to consider several aspects, including the quantity, duration, and timing of water diversion. Increasing outflow can reduce nutrient storage and phytoplankton biomass in lakes through shortening the water’s residence time [63]. In similar shallow lakes, such as Lake Taihu and Chaohu, multiple strategies have been implemented to improve lake water quality, including reduction of external loading from inflows, water transfer from the Yangtze River, increasing discharge in outflows, and sediment dredging in the heavily polluted area of the lake [64]. As a result, the water quality in both tributaries and water bodies was improved obviously in recent years [28].

4.3. Research Limitation and Prospects

It is worth noting that the sediment N and P release rates calculated in the present study have uncertainties. First, this estimate accounted for the static diffusive flux, and hence was likely lower than the actual flux, as we excluded bioturbation and related wind-disturbance fluxes. Previous studies have also claimed that internal N and P loads induced by resuspension were 8–10 times larger than the static release [65]. Second, the sediment characteristics in our site represented high levels of N and P in Lake Hulun [29], which may have resulted in an overestimation of the actual fluxes. Third, sediment nutrient release rates to a large extent depend on water trophic states [57]. Only using one-time sampling in an unfrozen period will inevitably cause errors, because the seasonal nutrient concentration in the overlying water of Lake Hulun is variable. Accordingly, our calculations likely represent high internal loading values, and provide an order of magnitude for the release fluxes from lake sediments in different water temperatures, DO and pH conditions (see Table 3). In Lake Tahoe, these data can still be used to preliminarily assess the impacts of internal nutrient loading on water column nutrient concentrations and algal productivity [66].

Our results demonstrate an urgent need for more systematic spatial and temporal monitoring of internal N and P release rates. In addition to the traditional lab incubation method used in this study, in situ approaches such as the benthic chamber can be used to measure solutes fluxes at the sediment–water interface [67]. In addition, ecological modelling could be used. For instance, Wang et al. (2023) built a sediment P cycling model and coupled it with the EcoLake model to simulate internal P loading throughout Lake Chaohu at different times of the year [4]. Due to the complexity of Lake Hulun and to climate change over the past several decades, further investigation of the spatiotemporal internal N and P loading and cycling in Lake Hulun is warranted.

5. Conclusions

Based on sediment incubations, we estimated the relatively large internal release rates of N and P from sediments to the water column exposed to potential environmental scenarios in Lake Hulun. The results obtained in this study indicate that: (1) internal N and P loading estimates for the whole lake accounted for 127% and 24% of external N and P loads; (2) the sediment ammonia and phosphate release rates in anoxic conditions were around three times higher than those in oxic conditions when other conditions were kept similar; (3) raised temperature caused higher phosphate release rates than ammonia. The highest temperature led to a slight decrease in the sediment ammonia release rate due to enhanced nitrification. Nitrate accumulated in the water column, and this was consistent with ammonia change, which verified strong biological nitrification in the lake water; (4) the pH increasing from 8.4 to 9.4 caused a significant decrease in sediment N and P release rates, a result which is opposite that of previous reports; (5) when comparing nutrient release rates between each treatment, both DO and pH are more influential factors affecting internal N and P loading, in contrast to temperature in Lake Hulun. Regarding the high internal nutrient loading in this lake, reducing the external N and P inputs from inflows and hay, implementing water diversion, and increasing outflow were suggested to improve water quality in Lake Hulun. These results significantly improved our understanding of internal N and P releases and the influential mechanisms behind simulated scenarios in plateau lake management. Due to the limited number of sampling sites and sampling times, the sediment N and P release rates obtained in the present study have uncertainties. In the future, routine monitoring of internal N and P release rates in different sampling sites (combined with ecological modelling) is needed to improve our understanding of the mechanisms of internal nutrient loading under different environmental circumstances.