The Influence of Seasonal Variability of Eutrophication Indicators on Carbon Dioxide and Methane Diffusive Emissions in the Largest Shallow Urban Lake in China

Abstract

Eutrophication is prevalent in urban lakes; however, a knowledge gap exists regarding eutrophication influences on carbon dynamics in these ecosystems. In the present study, we investigated the carbon dioxide (CO₂) and methane (CH₄) concentration and diffusion fluxes in Lake Tangxun (the largest shallow Chinese urban lake) in the autumn and winter of 2022 and spring and summer of 2023. We found that Lake Tangxun served as a source of GHGs, with average emission rates of 5.52 ± 12.16 mmol CO₂ m⁻² d⁻¹ and 0.83 ± 2.81 mmol CH₄ m⁻² d⁻¹, respectively. The partial pressure of dissolved CO₂ (pCO₂) (averaging 1321.39 ± 1614.63 μatm) and dissolved CH₄ (dCH₄) (averaging 4.29 ± 13.71 μmol L⁻¹) exceeded saturation levels. Seasonal variability was observed in the pCO₂ and dCH₄ as well as CH₄ fluxes, while the CO₂ flux remained constant. The mean pCO₂ and dCH₄, as well as carbon emissions, were generally higher in summer and spring. pCO₂ and dCH₄ levels were significantly related to total nitrogen (TN), total phosphorus (TP), and ammonium-nitrogen (N-NH₄⁺), and N-NH₄⁺ was a main influencing factor of pCO₂ and dCH₄ in urban eutrophic lakes. The positive relationships of pCO₂, dCH₄ and trophic state index highlighted that eutrophication could elevate CO₂ and CH₄ emissions from the lake. This study highlights the fact that eutrophication can significantly increase carbon emissions in shallow urban lakes and that urban lakes are substantial contributors to the global carbon budget.

1. Introduction

Lakes are an important component of inland water ecosystems and also hotspots of greenhouse gas (GHG), i.e., CO₂ and CH₄, release due to their high productivity and the exchange of material, energy, and information with terrestrial ecosystems [1,2,3,4,5]. Lacustrine CO₂ emission range from 0.11–0.57 Pg C yr⁻¹ [1,2,6,7]; likewise, lacustrine CH₄ evasion estimates vary even more widely, from 6 to 185 Tg CH₄ yr⁻¹ [8,9,10,11]. These research efforts underscore lakes as focal points for global carbon and CH₄ budgets, urging the need for meticulous investigation of lakes as a significant carbon source.

Urban lakes (normally with a mean depth of less than 3 m) are commonly shallow [12,13] and are indispensable components of landscape features and urban living environments [14,15]. Urban lakes have a vulnerability to a host of anthropogenic and environmental pressures and possess a poor capacity for self-purification [16]. Human activities, for instance, discharges of wastewater, can enhance loadings of organic carbon and nutrients in urban lakes, which can stimulate mineralization and thus promote methanogenesis [17]. Conversely, a surfeit of nutrients also enhances primary production, thereby increasing a lake’s carbon sequestration capacity, and thus concurrently curtailing CO₂ emissions [18,19]. Beyond nutrients, environmental factors such as wind and temperature also clearly affect urban lakes. Wind-induced turbulence, for instance, can reduce stratification and mitigate hypoxic conditions, thereby augmenting benthic oxygen availability and restraining CH₄ production, resulting in reduced CH₄ emissions and increasing CO₂ emissions [20,21]. Temperature, however, plays a multifaceted role, fostering primary production, which in turn can reduce CO₂ production [18,19]. However, studies exploring the characteristics and influencing factors of GHG emissions from urban lakes remain relatively scarce in the literature [22,23,24].

Eutrophication occurs widely in lakes and exerts a pivotal role in regulating GHG emissions from lakes [19,25,26]. It is worth noting that anthropogenic eutrophication is a predominant driver of lake eutrophication [27]. Studies have shown that eutrophic lakes can substantially enhance CH₄ production and emission [28,29]; however, the findings regarding CO₂ production and emission in eutrophic lakes are divergent. Some studies have demonstrated CO₂ emissions from eutrophic lakes to the atmosphere [25], yet others have reported a decrease in CO₂ emissions [19] or even a net CO₂ uptake from the atmosphere [30]. Notwithstanding the efforts made, the intricate web of CO₂ and CH₄ production and emission in response to lake eutrophication remains enigmatic and warrants further investigation. This knowledge gap represents the exigency for in-depth study to disentangle the nuances governing carbon in eutrophic lakes. As such, carbon exchange at the air–water interface, especially in eutrophic urban lakes, is an urgent imperative.

In the study, we used Lake Tangxun, the largest shallow urban lake located in a subtropical region, as an example to delve into the GHG dynamics of a eutrophic lake. The overarching objectives of our research endeavors were to (1) unravel the seasonal variation and driving factors in both the concentrations and fluxes of CO₂ and CH₄; (2) elucidate the profound effects of eutrophication on GHG dynamics in a shallow urban lake. We tested the hypotheses that (1) key drivers of seasonal fluctuations in GHGs concentrations and fluxes are variable, and (2) eutrophication considerably shapes the behavior of CO₂ and CH₄ in this specific lake ecosystem. The study aims to enhance our comprehension of GHG emissions in eutrophic shallow urban lakes, shedding light on informing effective mitigation strategy and management practice of lake carbon evasion, particularly within the context of escalating urbanization and its associated environmental challenges.

2. Materials and Methods

2.1. Study Area

Lake Tangxun (30°22′–30°30′ N, 114°15′–114°35′ E) is the largest Chinese urban lake. It lies in Wuhan, Hubei Province. Lake Tangxun includes an inner lake in the east, making up 33% of the total area, and an outer lake in the west (Figure 1). The water surface area and average depth of Lake Tangxun are 47.6 km² and 1.85 m, respectively [31,32]. Situated within a subtropical climatic zone, the lake is ice-free during winter months. The annual mean temperature and precipitation in the area are 18.3 °C (ranged −6–38 °C) and 1057.3 mm, respectively (Figure S1). The development of industry in the catchment area is a trajectory representing the onset of eutrophication in 1996 [31]. This evolution underscores the intricate interplay between urbanization environmental shifts, rendering Lake Tangxun a compelling natural laboratory for understanding eutrophication and its implications.

2.2. Sample Collection and Measurements

Field sampling and surveys of Lake Tangxun were diligently executed in October 2022 and February, April, and June 2023. These temporal selections align with autumn, winter, spring, and summer, respectively. Twenty strategically designated sampling sites were meticulously selected in Lake Tangxun (Figure 1). This ensures the representativeness of our data collection. Site 1 is an aquaculture pond. Sites 2–3, 9–12, 15–18, and 20 are located near the sewage outlet; meanwhile, sites 9, 12, and 16 are the non-point source pollution catchments, and site 15 is in a village. Sites 4–8 and 19 are in Wetland Park. Sites 13 and 14 are campsites; additionally, site 14 also exists within a small amount of cropland. There is no seasonal variation in wastewater discharge, while rainfall runoff experiences seasonal variations [33]. It is noteworthy that no extreme hydrometeorological events were recorded during each sampling campaign, and the sampling as a whole represented the prevailing climatic and hydrological conditions in the region.

A Multi-Parameter Meter (Thermo Fisher Eutech, Singapore) was used to record the in situ surface water temperature (T_W), pH, electrical conductivity (EC) and dissolved oxygen (DO), with accuracies of ±0.5 °C, ±0.002 pH units, ±1.0% and ±2.0%, respectively. A TES-1341 hot-wire anemometer (TES Corp., Taiwan, China) meter was used to record ambient temperature (T_a) and wind speed (U) about 1 m above overlying water, and a pressure probe recorded air pressure. Water samples (0.1–0.3 m beneath water surface) were gathered using a 2.5 L plexiglass hydrophore. Subsequently, water samples were filtered with pre-burnt glass fiber filter (0.7 μm pore size, Whatman, Maidstone, UK), and the filter membranes were used to measure chlorophyll a (Chl-a) via acetone extraction. Transparency (SD) was measured with a Secchi disk. All samples were placed in ice boxes in field and promptly preserved in refrigerators or subjected to freezing after reaching the lab.

Ultraviolet spectrophotometry (UV-8000, Shanghai Yuanxi Instrument Co., Ltd., Shanghai, China) can be applied to measure nutrient concentrations, including total nitrogen (TN), total phosphorus (TP), ammonium-nitrogen (N-NH₄⁺), and nitrate-nitrogen (N-NO₃⁻) (https://www.sac.gov.cn//, accessed on 1 November 2022). A TOC analyzer (Shimadzu, TOC-L, Kyoto, Japan) was used to analyze dissolved organic carbon (DOC).

The well-established headspace equilibrium method can directly measure CO₂ and CH₄ concentrations [34,35,36]. Seventy-mL water samples (0.1 m beneath water surface) were extracted via a 100 mL gas-tight syringe; subsequently a 30 mL volume of ambient atmosphere was extracted for equilibrating. The syringe was vigorously shaken to accelerate equilibrium. The resulting headspace gas samples were collected in a vacuum aluminum foil airbag (Delin Gas Packing Co., Ltd., Dalian, China). Meanwhile, ambient air was extracted at the overlying water. Gas samplings ware analyzed with a gas chromatograph (Shimadzu, GC-2014, Kyoto, Japan). CO₂ and CH₄ concentrations were computed by Henry’s law as follows [36]:

pCO₂ = C_G × R × T + V_g/V_w × (C_g-C − C_a-C) × K_H-C

(1)

dCH₄ = C_g-M × K_H-M + V_g/V_w × (C_g-M − C_a-M)

(2)

ln K_H-C = −58.0931 + 90.5069 × (100/T) + 22.2940 × ln (T/100)

(3)

K_H-M = β/22.356

(4)

ln β = −68.8862 + 101.4956 × (100/T) + 28.7314 × ln (T/100)

(5)

where pCO₂ is partial pressure of dissolved CO₂ (μatm); dCH₄ is dissolved CH₄ (μmol L⁻¹); C_g-M (or C_g-C) is the CH₄ (or CO₂) concentration of gas samples (μmol L⁻¹); T is measured T_W (K); R is prevalent gas constant (0.082057 L atm mol⁻¹ K⁻¹); K_H-M (or K_H-C) is Henry solubility constant (mol L⁻¹ atm⁻¹) of CH₄ (or CO₂); V_g is the air volume and V_w is the water volume in balance (mL); C_a-M (or C_a-C) is atmospheric concentration of CH₄ (or CO₂) (μmol L⁻¹).

Theoretical diffusion model was used to calculate the gas diffusive fluxes (F, mmol m⁻² d⁻¹) as follows [37,38]:

F = k × (C_w − C_a)

(6)

k = k₆₀₀ × (S_c/600)⁻ⁿ

(7)

S_c(CO₂) = 1911.1 − 118.11 × t + 3.4527 × t² − 0.04132 × t³

(8)

S_c(CH₄) = 1897.8 − 114.28 × t + 3.2902 × t² − 0.039061 × t³

(9)

where k is the gas piston coefficient (cm h⁻¹); (C_w − C_a) is concentration gradient (μmol L⁻¹). S_c is the Schmidt number regulated by water temperature (t, °C), the exponent n is a constant determined by measured wand speed, k₆₀₀ is calculated using an empirical formula [39].

To elucidate the trophic status of Lake Tangxun and its effect on GHG dynamics, four major parameters were applied to calculate the trophic state index (TSI). TSI is calculated as follows [18]:

TSI = 0.219TSI(TN) + 0.230TSI(TP) + 0.326TSI(Chl-a) + 0.225TSI(SDD)

(10)

TSI(TN) = 10 × (5.453 + 1.694 × lnTN)

(11)

TSI(TP) = 10 × (9.436 + 1.624 × lnTP)

(12)

TSI(Chl-a) = 10 × (2.5 + 1.086 × lnChl-a)

(13)

TSI(SDD) = 10 × (5.118 − 1.94 × lnSD)

(14)

where TSI values of 30–50, 50–60, 60–70 and >70 represent mesotrophic, eutrophic, moderately eutrophic and highly eutrophic, respectively.

2.3. Statistical Analyses

The substantial portion of the data were non-normal. The significant differences of the GHG concentrations and emissions, environmental parameters amongst seasons were examined via Kruskal–Wallis test. A significance probability of p < 0.05 was used. Non-normal data are logarithmically transformed when necessary. Pearson’s correlation could explore relationships of gas concentration and physiochemical factors. IBM SPSS statistics 26 and Origin 2021 were used for statistical analyses. Origin 2021 was used for drawing figures.

3. Results

3.1. Environment Factors

Lake Tangxun showed substantial seasonal differences in physicochemical and biological properties (Figure 2). There were statistical differences in T_W, transparency, DO, EC, Chl-a, DOC, and N-NO₃⁻ concentrations among seasons (p < 0.001); nevertheless, no statistical differences existed in U, pH, N-NH₄⁺, TN or TP across seasons. The average T_W in winter (11.77 ± 1.53 °C) was significantly lower than in spring (26.28 ± 2.04 °C), summer (27.01 ± 1.55 °C), and autumn (21.04 ± 1.20 °C) (Figure 2a). The highest and lowest average of U and SD appeared in winter (1.33 ± 1.45 m s⁻¹ and 0.63 ± 0.22 m) and summer (0.76 ± 0.93 m s⁻¹ and 0.35 ± 0.16 m), respectively (Figure 2b,c).

The average pH value was slightly higher in autumn (8.14 ± 0.59) compared to the other seasons (Figure 2d). The DO concentrations and EC values firstly decreased and then increased (Figure 2e,f). They both were highest in winter (9.90 ± 1.45 mg L⁻¹ for DO and 522.60 ± 109.51 μS cm⁻¹ for EC, respectively) and lowest in summer (6.21 ± 2.27 mg L⁻¹ for DO and 394.00 ± 75.13 μS cm⁻¹ for EC, respectively). Inversely, the highest average Chl-a concentrations appeared in summer (125.81 ± 103.13 μg L⁻¹) and lowest appeared winter (27.78 ± 16.72 μg L⁻¹) (Figure 2g), respectively.

The mean concentrations of DOC, N-NH₄⁺, TN, and TP first increased and then decreased (Figure 2h,i,k,l). The highest DOC (5.64 ± 0.93 mg L⁻¹), N-NH₄⁺ (2.47 ± 5.62 mg L⁻¹), TN (4.25 ± 5.75 mg L⁻¹) and TP (0.31 ± 0.58 mg L⁻¹) were found in autumn but were lowest in spring (4.11 ± 0.99 mg L⁻¹, 1.44 ± 1.90 mg L⁻¹, 3.15 ± 2.56 mg L⁻¹ and 0.20 ± 0.14 mg L⁻¹, respectively). The highest average N-NO₃⁻ concentration was in winter (0.87 ± 0.64 mg L⁻¹) and lowest in summer (0.35 ± 0.38 mg L⁻¹) (Figure 2j).

The TSI of Lake Tangxun showed a statistically significant difference among seasons (p < 0.01). The average TSI (67.68 ± 8.25) was less than 70, illustrating that the Lake Tangxun was a medial eutrophic lake overall (Figure 3). The pronounced seasonality in the TSI unveiled a highest value during the summer season (averaging 72.04 ± 6.77), but the lowest during the winter season (averaging 63.06 ± 6.91). Lake Tangxun was hypereutrophic in the summer, while it transitioned to a state of moderate eutrophication in the other seasons.

3.2. pCO₂ and CO₂ Flux

The mean pCO₂ is 1321.39 ± 1614.63 μatm (range: 86–10,968 μatm). Of the measured pCO₂ data, 85% exceeded the atmospheric CO₂ average (526 μatm). The average pCO₂ in spring (1621.50 ± 1387.83 μatm) and summer (2217.91 ± 2641.93 μatm) was significantly higher than that in autumn (647.71 ± 172.21 μatm) and in winter (798.45 ± 444.68 μatm) (p < 0.05) (Figure 4a).

The mean CO₂ flux is 5.52 ± 12.16 mmol m⁻² d⁻¹ (range: −15.83 – 67.30 mmol m⁻² d⁻¹). There was no significant difference amongst seasons (Figure 4b). The average CO₂ flux followed a decline trend as follows: spring (9.36 ± 13.10 mmol m⁻² d⁻¹) > summer (6.49 ± 16.88 mmol m⁻² d⁻¹) > winter (4.39 ± 10.17 mmol m⁻² d⁻¹) > autumn (1.85 ± 4.81 mmol m⁻² d⁻¹).

There were significant negative relationships of pCO₂ with pH, DO and Chl-a (Figure 5a–c). Additionally, pCO₂ had a strong positively relationship with N-NH₄⁺, TN, TP, and TSI (Figure 5d,f and Figure 6a).

3.3. dCH₄ and CH₄ Flux

The overall dCH₄ changed between 0.01 and 102.00 μmol L⁻¹ (mean: 4.29 ± 13.71 μmol L⁻¹). The mean value of atmospheric CH₄ concentration was 0.12 ± 0.01 μmol L⁻¹, illustrating oversaturation of CH₄ in Lake Tangxun. The mean dCH₄ in autumn (0.83 ± 2.90 μmol L⁻¹) was lower than that in other seasons (spring: 5.74 ± 11.01 μmol L⁻¹, summer: 6.25 ± 22.58 μmol L⁻¹ and winter: 4.33 ± 11.10 μmol L⁻¹) (p < 0.001) (Figure 4c).

The overall CH₄ fluxes ranged between 5.48 × 10⁻⁵ and 20.42 mmol m⁻² d⁻¹, with a grand average of 0.83 ± 2.81 mmol m⁻² d⁻¹. The average CH₄ flux followed a decline trend as follows: spring (1.51 ± 3.06 mmol m⁻² d⁻¹) was slightly higher than summer (1.21 ± 4.53 mmol m⁻² d⁻¹) and winter (0.46 ± 1.30 mmol m⁻² d⁻¹), and significantly higher than autumn (0.13 ± 0.25 mmol m⁻² d⁻¹) (p < 0.05) (Figure 4d).

dCH₄ exhibited significantly negative correlations with pH and DO (Figure 7a,b), while showed significantly positive correlations with N-NH₄⁺, TN and TP (Figure 7c–e). dCH₄ also revealed a significantly positive correlations with pCO₂ (Figure 7f) and TSI (Figure 6b).

4. Discussion

4.1. Seasonal and Spatial Shifts of CO₂ and CH₄

We found clearly seasonal variations in pCO₂ with significantly higher average pCO₂ in spring and summer in Lake Tangxun (p < 0.05; Figure 4). This was attributable to the higher temperature and suitable precipitation in the two seasons (Figure S1). Higher temperature can increase in-site microorganism metabolism, and high temperature and precipitation promote heterotrophic respiration in the littoral soils; thus the surrounding soils have a higher CO₂ content, thereby transporting more CO₂ to lakes via runoff [34,40,41,42]. This was supported by the positive correlations of pCO₂ with T_a (p < 0.01) and precipitation (p < 0.01) (Figure S2).

Moreover, nutrient variables can favor in-stream respiration of DOC and photosynthetic up-taking of CO₂; this process can be responsible for the positive relationships between pCO₂ and nutrient elements of TN, TP and N-NH₄⁺, and the negative relationship between pCO₂ and Chl-a (Figure 5) [15,43].

The much lower dCH₄ observed in autumn (Figure 4c) was contrary to previous studies [19,44,45]. The authors ascribed this seasonal pattern to the synthetic effect of temperature and organic carbon availability [19,45,46]. Warmer temperatures can promote methanogenesis by affecting microbial metabolism [47], which can be confirmed by the remarked relationship between T_w and dCH₄ (Table S2). This can be ascribed to the faster decomposition rate of organic matter and consumption of dissolved oxygen at high temperatures, creating favorable conditions for methanogenesis, whilst frequent rain events in spring and summer (Figure S1) could potentially transport soil CH₄ into lakes and promote CH₄ diffusion from bottom to surface in lakes [48,49]. Moreover, although the highest mean DOC concentration occurred in autumn, internal respiration of organic carbon in autumn was lower than in winter, which can be confirmed by the correlation of DOC and dCH₄ (Table S2), resulting in dCH₄ in winter higher than that in autumn. We need to stress that seasonal variation of CH₄ dynamics was relatively complex in urban lakes.

We found that some points are significantly deviated from the average values (Figure 4), suggesting huge variability of data. The deviated values of pCO₂, dCH₄, CO_2, and CH₄ diffusion fluxes are highly related to the environment characteristics of sampling points. The pCO₂ and dCH₄ were highly related to N-NH₄⁺, N-NO₃⁻, and Chl-a (

Table 1. Stepwise regression with pCO₂ and dCH₄ and ambient conditions in Lake Tangxun. R² is adjust R².

	CO2	CH4
Multiple regression	pCO2 = 2.61 + 0.29N-NH4+ − 0.23N-NO3− − 0.36Chl-a + 0.72Tw	dCH4 = 2.00 + 0.41N-NH4+ − 0.89N-NO3− − 0.66Chl-a − 2.85DOC + 1.40TN
R2	0.44	0.59
p	<0.001	<0.001
Part R2	0.27 ***, 0.05 *, 0.07 **, 0.05 **	0.40 *, 0.07 ***, 0.06 ***, 0.02 **, 0.04 **

Note: *, ** and *** represent p < 0.05, p < 0.01 and p < 0.001.

). The sites with deviated values of pCO₂ and dCH₄ exhibited higher N-NH₄⁺; for instance, the highest concentrations of N-NH₄⁺ and gases simultaneously occurred at site 20 (Table S1). CO₂ and CH₄ diffusion fluxes were not dependent on pCO₂ and dCH₄ but closely related to U (Figure S3). Random measures of wind speed could lead to CO₂ and CH₄ diffusion fluxes deviated from the average values at those sites with high wind speed.

4.2. pCO₂ Control Factors

In-stream biogeochemical processes mainly control lake pCO₂ [50]. Chl-a can be a proxy of lacustrine primary productivity; thus, increasing Chl-a concentration can promote photosynthesis [29,30,34]. This photosynthetic process produces DO but consumes CO₂, thereby resulting in the significant negative relationships of pCO₂ with Chl-a (Figure 5c), DO (Figure 5b), and pH (Figure 5a). Previous studies support the reciprocal relationships [26,51].

Nutrient loading can influence lake photosynthesis and bacterial respiration, thereby regulating pCO₂ in lakes [18,19]. For example, nutrients can favor both photosynthesis and in-site respiration of carbon. The decay of algae can add organic matter to the sediment and stimulate endogenous respiration in lakes [34]. In our studied lakes, lake-dissolved CO₂ is over-saturated, suggesting that the studied lakes are heterotrophic, thereby leading to the relationships of pCO₂ with N-NH₄⁺, TN, and TP (Figure 5d–f). The stepwise regression analysis model showed that N-NH₄⁺ was a good predictor for estimating pCO₂ in Lake Tangxun (Table 1). N-NH₄⁺ is a good proxy for sewage loads in the lakes [14], implying the urban sewage contribution of high pCO₂.

We found that pCO₂ increased as eutrophication status increased (Figure 6a). The effect of eutrophication on CO₂ was closely bound up the interaction of CO₂ consumption and production [26,52]. Previous studies have found algae blooms enabled the lake to absorb CO₂ from the atmosphere [53,54]. We, however, observed an atmospheric CO₂ source of the studied lake even in the summer season, with much higher Chl-a concentration. This implied the importance of respiration rather than photosynthesis in the eutrophic Lake Tangxun [55,56]. Thus, Lake Tangxun, a net CO₂ source, is characterized by a heterotrophic nature, supporting the positive associations between pCO₂ and nutrient species (Figure 5).

4.3. dCH₄ Control Factors

Geochemical controls of CH₄ dynamics in aquatic ecosystems are complicated [37,57]. First, CH₄ production and consumption are controlled by methanogenic and aerobic methane-oxidizing bacteria [37,58,59]. They are also affected by algae decay in eutrophic lakes [46,60,61]. Previous studies have also showed that algal blooms have a catalytic effect on CH₄ dynamics [53,62]. Second, variations in microenvironmental factors such as pH and DO can affect CH₄ production and consumption [45,63]. CH₄ is easily oxidized to carbon dioxide; thus, low DO concentration can decrease CH₄ oxidation and increase CH₄ production [64,65]. These processes led to the negative correlation between dCH₄ and DO (Figure 7b). The positive correlation between dCH₄ and pCO₂ (Figure 7f) can support CH₄ oxidation to CO₂, further evidencing the association between dCH₄ and DO. Similar results from lakes in Northeastern China and Lake Kivu in East Africa support our findings [66,67]. It was interesting that we also found a significantly negative correlation between dCH₄ and pH (Figure 7a). This can be explained by the pH range of 7 to 9 in Lake Tangxun. Prior studies showed that pH ranging between 6 and 8 is the most favorable for methanogenesis, and CH₄ concentrations decline as pH values drop out of the optimal range [63]. For instance, dCH₄ was minimal in autumn (Figure 4c) when an average pH (8.14 ± 0.59) was outside the optimal range for CH₄ production (Figure 2d,h).

The content, availability, and redox conditions of organic carbon are the pivotal factors affecting CH₄ dynamics [46]. In this study, multiple nutrients (N-NH₄⁺, TN and TP) regulated the dCH₄ in Lake Tangxun (Figure 7c–e). The input of nutrients can be used as the basic substrate in methanogenesis [19,46]. Meanwhile, nitrogen input also promoted CH₄ production by providing sufficient substrate for the methanogens’ growth [57,68]. N-NH₄⁺ was also a good proxy for estimating dCH₄ in Lake Tangxun (Table 1). It proved ammonia nitrogen can prevent CH₄ oxidation [69], as reflected by the correlation between N-NH₄⁺ and dCH₄ (Figure 7c).

Consistent with pCO₂, eutrophication increased dCH₄, as suggested by the significantly positive correlation between TSI and dCH₄ (Figure 6b). Several studies supported our findings [19,29,62]. Eutrophic lakes provided most labile organic matter that can be promptly converted to CH₄ precursors [28], as reflected by the positive correlation between dCH₄ and nutrients (Figure 7c–e). Beaulieu and DelSontro [62] also found enhanced eutrophication of lentic waters could substantially increase CH₄ emission. Meanwhile, shallow eutrophic lakes are suitable for submerged vegetation, which can also provide substrates to promote CH₄ production [70]. Moreover, Sepulveda-Jauregui and Hoyos-Santillan [71] found that eutrophic lakes had higher temperature sensitivity, resulting in higher CH₄ emission. In Lake Tangxun, in summer, the TSI and temperature were the highest, and dCH₄ was the highest (Figure 3 and Figure 4c). In winter, although there was slightly higher DOC concentration than in spring and summer, the trophic status and temperature was the lowest, leading to relatively lower dCH₄.

4.4. Implications of CO₂ and CH₄ Flux

Urban lakes had notable spatial–temporal heterogeneity of CO₂ flux, and acted as a considerable atmospheric CO₂ source (Table 2). CO₂ flux (averaging 5.52 mmol m⁻² d⁻¹) in this study was lower than in other studies (averaging 20.80 mmol m⁻² d⁻¹) (Table 2). Sydney’s severe eutrophic urban lake showed the peaking flux of CO₂ [72]. What is noteworthy is that CO₂ emissions were not only subject to lake nutrient loadings, but also controlled by seasonal variation of ambient conditions, as exhibited in recent studies in subtropic urban eutrophic lakes [54,73]. Sun et al. [19] and Zhang et al. [54] found eutrophic lakes converted to be CO₂ sinks in summer due to higher primary productivity, which was different from our results in Lake Tangxun. This implied that seasonal sampling can lead to large uncertainty of lake CO₂ emission.

Urban lakes in subtropical zones showed higher CH₄ fluxes than those in boreal zones (Table 2). This was likely due to the rate of methanogenesis and substrate production increase with a warmer temperature [74,75,76]. Furthermore, inputs of sewage also contributed to higher CH₄ fluxes in urban lakes, such as the severely polluted Lake Bellandur [22]. Besides the study lake ecosystem, other aquatic ecosystems in urban areas also showed higher CH₄ fluxes [5,36,77]. Rosentreter, Borges [11] also demonstrated that CH₄ emissions are likely to increase because of eutrophication and urbanization based on a metadata analysis of CH₄ fluxes in all major aquatic ecosystems. We suggest that more field measures should be conducted, to better understand CH₄ dynamics in lakes.

Table 2. CO₂ and CH₄ diffusive fluxes in others urban lakes in different climate zones.

Lake	City	Climate Zone	Sampling Time	FCO2 (mmol m−2 d−1)	FCH4 (mmol m−2 d−1)	Source
Lake Vesijärvi	Finland	Boreal	May–October 2018	12.4 ± 2.38	0.24 ± 0.06	[24]
Lake Obersee, et al.	Berlin	Temperate	April-May and July–October 2016, February–March 2017		2.19	[76]
Sydney’s largest urban lake	Sydeney	Temperate	June and July 2019	113 ± 81	0.3 ± 0.1	[72]
Lake Lyng	Silkeborg	Temperate	September–October 2021	22.05	1.68	[78]
Lake SCH and YYT	Beijing	Temperate	July 2018–November 2019	−0.2 ± 13.0	0.7 ± 0.6	[79]
Lake Wuli	Wuxi	Subtropic	2000–2015	25.0 ± 13.64		[14]
			2002–2017	21.12 ± 19.60		[15]
Lake Donghu	Wuhan	Subtropic	April 2003–March 2004	7.7	1.6	[45]
			2002–2017	16.42 ± 20.39		[15]
Lake Nanhu, et al.	Wuhan	Subtropic	October–December 2021, February–March and May–June 2022	3.65	1.32 ± 4.11	[54,57]
Lake Tangxun	Wuhan	Subtropic	October 2022, February, April and June 2023	5.52 ± 12.16	0.83 ± 2.81	This study
Lake Bhalswa	Dehli	Tropic	Summer 2018	11.65 ± 3.42		[80]
			Winter 2017	6.33 ± 2.23
Lake Bellandur	Bangalore	Tropic	June 2018–February 2020	5.81	3.6	[22]
Lake Jakkur	Bangalore	Tropic	June 2018–February 2020	4.5	1.48	[22]

Table 2. CO₂ and CH₄ diffusive fluxes in others urban lakes in different climate zones.

Lake	City	Climate Zone	Sampling Time	FCO2 (mmol m−2 d−1)	FCH4 (mmol m−2 d−1)	Source
Lake Vesijärvi	Finland	Boreal	May–October 2018	12.4 ± 2.38	0.24 ± 0.06	[24]
Lake Obersee, et al.	Berlin	Temperate	April-May and July–October 2016, February–March 2017		2.19	[76]
Sydney’s largest urban lake	Sydeney	Temperate	June and July 2019	113 ± 81	0.3 ± 0.1	[72]
Lake Lyng	Silkeborg	Temperate	September–October 2021	22.05	1.68	[78]
Lake SCH and YYT	Beijing	Temperate	July 2018–November 2019	−0.2 ± 13.0	0.7 ± 0.6	[79]
Lake Wuli	Wuxi	Subtropic	2000–2015	25.0 ± 13.64		[14]
			2002–2017	21.12 ± 19.60		[15]
Lake Donghu	Wuhan	Subtropic	April 2003–March 2004	7.7	1.6	[45]
			2002–2017	16.42 ± 20.39		[15]
Lake Nanhu, et al.	Wuhan	Subtropic	October–December 2021, February–March and May–June 2022	3.65	1.32 ± 4.11	[54,57]
Lake Tangxun	Wuhan	Subtropic	October 2022, February, April and June 2023	5.52 ± 12.16	0.83 ± 2.81	This study
Lake Bhalswa	Dehli	Tropic	Summer 2018	11.65 ± 3.42		[80]
			Winter 2017	6.33 ± 2.23
Lake Bellandur	Bangalore	Tropic	June 2018–February 2020	5.81	3.6	[22]
Lake Jakkur	Bangalore	Tropic	June 2018–February 2020	4.5	1.48	[22]

The eutrophication effect of GHG emissions in the urban lakes was not fully understood. Our results provided new understanding into seasonal fluctuations of CO₂ and CH₄ diffusive fluxes in a middle eutrophic urban lake. We highlighted that eutrophication could significantly promote CO₂ and CH₄ from lake into atmosphere. This is particularly evident in the eutrophic lakes of the Yangtze Basin, as supported by increasing lake CO₂ and CH₄ emissions as the trophic state increase [15,57]. Hence, the management (e.g., retreatment of wastewater) and ecological restoration (e.g., sediment dredging) measures of lakes are necessary in highly urban-dominated regions, which could not only decrease lacustrine carbon emission in urban lakes, but also improve water quality, further contributing to the goal of carbon neutrality.

5. Conclusions

The study investigated the seasonal dynamics and potential control factors of dissolved CO₂ and CH₄ in the largest Chinese urban Lake. Our results illustrated that Lake Tangxun was atmospheric greenhouse gases (GHGs) source. CO₂ and CH₄ concentrations and thus diffusion fluxes significantly increased as trophic status index increased. We further found that N-NH₄⁺ was a predictor of pCO₂ and dCH₄. More studies should be conducted to accurately quantify carbon emission in eutrophic lakes, particularly in urban areas that are intensively affected by human activities. The study highlighted the importance of eutrophic control for restoration of water environment and reducing CO₂ and CH₄ emissions. Meanwhile, this study also stressed that urban lakes, which are heavily influenced by human activities, exhibited more complex seasonal dynamics in CO₂ and CH₄ fluxes. Therefore, urban lakes should be considered an important component of carbon budgets.