Invasive Water Hyacinth (Eichhornia crassipes) Increases Methane Emissions from a Subtropical Lake in the Yangtze River in China

Abstract

Lakes represent an important source of atmospheric methane (CH₄); however, there are few studies on which lake-dwelling invasive aquatic plants generate CH₄. Therefore, in this study, CH₄ emissions were measured using a floating chamber and gas chromatography in a subtropical lake in China. We considered four community zones of invasive plants (Eichhornia crassipes), emergent vegetation (Zizania latifolia), floating-plant (Trapa natans) and open-water zones. The results indicate that the flux of CH₄ emissions varied between −5.38 and 102.68 mg m⁻² h⁻¹. The higher emission values were attributed to lake eutrophication. Moreover, the flux of CH₄ emissions in the invasive plant zone was 140–220% higher than that in the open-water and the floating-plant zones. However, there was no significant difference in CH₄ emissions between the invasive plant and the emergent vegetation zones. This may be due to a higher production of plants, as well as the rapid reproductive rate of the invasive plants. Finally, CH₄ emissions were positively associated with the air and water temperature; however, the emissions were also negatively associated with water depth. Our results suggest that invasive plants enhance freshwater CH₄ emissions, thus contributing to global warming.

1. Introduction

Despite the fact that lake ecosystems cover approximately 3.7% of the Earth’s continental land area [1,2], they are believed to be a major source of the greenhouse gas (GHG) methane (CH₄) [3,4]. The greenhouse effect of CH₄ is approximately 28-fold higher than that of CO₂ for the century-long time scale, accounting for approximately 20% of total global warming [3,5]. The CH₄ concentration in the atmosphere, which is mainly caused by human activity, has increased by 150% since pre-industrial times and continues to increase. This may further enhance global warming to a greater extent [5,6].

Top-down and bottom-up estimates for global CH₄ emissions are 576 and 727 Tg yr⁻¹, respectively, of which CH₄ emissions from freshwater wetlands (including lakes and rivers) account for 308 Tg yr⁻¹ [5]. CH₄ emissions between and within lakes exhibit high spatiotemporal variability [7,8], and it is estimated that CH₄ emissions from lakes account for 8–48 Tg yr⁻¹, with approximately 50% of the flux being attributed to tropical/subtropical regions [3,9]. Although several studies have determined the CH₄ emissions from these lakes are largely the result of a warming climate, invasive alien plants, pollution and enclosure aquaculture [10,11,12,13], there are few studies on the contribution of human activity (e.g., introduction of alien plants) to CH₄ emissions from lakes. In addition, these studies suggest that current and future increases in CH₄ emissions will intensify climate change [12,13,14]; therefore, it is necessary to further explore CH₄ emissions from lakes.

The free-floating water hyacinth (Eichhornia crassipes) is one of the world’s most invasive aquatic plants. It causes significant ecological and socio-economic effects [15]. As an ornamental plant originating from tropical South America, this invasive water hyacinth weed was introduced into China in the 1900s [16], and it has subsequently been extensively cultivated as animal feed since the 1950s. It is distributed widely in the aquatic ecosystems of the Yangtze River in China [16]. The water hyacinth commonly forms dense, interlocking mat-forming floating aquatic plants on the water surface. This results from a rapid reproductive rate, complex root structure and a doubling of its biomass within five days [10,15,17,18]. The water hyacinth mats prevent the transfer of oxygen from the air to the water’s surface and block the light required for photosynthesis by phytoplankton and submersed vegetation [15,18]. The water hyacinths on the water surface can prevent light penetration into the water column below [18], which decreases the temperature (water and sediment) [19,20]. Finally, changes in these factors affect the spatiotemporal variability of CH₄ production and emissions from the lake and impact whole-lake emission estimates on an annual basis [3,7,10]. Therefore, it is important to study the influence of the invasive water hyacinth weed on CH₄ emissions in these lakes.

Hong Lake is the seventh largest shallow lake in China and the largest natural lake in Hubei Province, which is located within the middle reaches of the Yangtze River [21]. Because of the abundant natural resources in Hong Lake, rapid socio-economic development has caused the lake to undergo a variety of environmental changes over the past few decades, including a shrinking water area, deterioration of water quality and a decline in biodiversity [21,22,23]. Several studies have confirmed that CH₄ emissions from the invasive Spartina altrniflora weed have significantly increased compared with the CH₄ emissions from the native plant community [24,25,26]. Banik et al. [27] also reported that the invasive water hyacinth has clearly increased CH₄ emissions from the freshwater ecosystems in India, which were estimated to reach 1.2 Tg yr⁻¹. Conversely, Attermeyer et al. [10] reported that CH₄ emissions from invasive water hyacinth zones were reduced compared with those from open-water zones. This was caused by the oxidation of CH₄ catalysed by methanotrophic bacteria. Therefore, the influence of invasive aquatic plants on CH₄ emission rates in freshwater ecosystems requires further examination.

In the present study, we focused on the effects of the invasive water hyacinth on CH₄ emissions in a shallow lake in subtropical China. CH₄ emissions in the shallow lake were measured with floating chambers and a gas chromatography method. In addition, to reveal the scope of its influence, we analysed the relationship between ecological factors (water depth, temperature and dissolved oxygen from water) and CH₄ emissions.

2. Materials and Methods

2.1. Study area Description

The study was conducted at the Hong Lake Natural Reserve (113°12′–113°26′ N, 29°40′–29°58′ E) towards the middle reaches of the Yangtze River. Hong Lake has a surface area of 344 km² with an open-water area of 308 km², a littoral area of 36 km² and a mean water depth of 1.5 m [21,28]. Hong Lake was listed as an internationally important wetland in the Ramsar convention in 2008 and was entered into the China Wetland Ecosystem Research Network in 2014. The region is characterised by a north subtropical humid monsoon climate, with a mean annual temperature of 15.9 °C–16.6 °C. The minimum and maximum mean monthly temperatures were 3.8 °C in January and 28.9 °C in July and/or August, respectively. The annual average evaporation is 1000–1300 mm and the mean annual precipitation is 1174 mm, 74% of which occurs between April and October [21].

According to the change in water depth and the vegetation type from the littoral zone to the open water, four zones in the study region were selected to monitor CH₄ emission flux between April and October of 2021 (Figure 1). The first site (OPs) was the open water of the lake in which no vegetation grew. The second site (ECs) was an area invaded by an alien species of water hyacinth (E. crassipes), which covered 100% of the area. The third site (TNs) was covered by the floating plant Trapa natans (T. natans), with a total vegetation coverage of 95%. The fourth site (ZLs) consisted of the emergent aquatic plant Zizania latifolia (Z. Latifolia), with a total vegetation coverage of 90% and sparse areas containing Nelumbo nucifera and T. natans.

2.2. CH₄ Measurements

CH₄ flux measurements were taken at four sites in Hong Lake from April to December 2021. The measurements were carried out using floating chambers [29], which included three plastic opaque chambers (height above the water level 30 cm, volume 28.8 L) made of acrylic organic glass. The outside of the chamber contained a rubber plastic film to prevent an increase in the inner temperature of the chamber (Figure 2). In addition, the open-end of the chamber was fitted with a cystosepiment and tyre as floating equipment. The headspace of the chamber was equipped with a fan to mix the air and one sampling port and a temperature sensor. Before sampling, three chambers were placed upside down 50 to 100 cm apart on the water surface. Gas samples were drawn from each chamber every 5 min for 15 min with 60 mL polypropylene syringes equipped with three-way stopcocks and then transferred to a gas bag.

Within one week, the CH₄ concentration of all samples was determined by a gas chromatography instrument (Agilent, 7890A, GC system, Agilent Co., Wilmington, DE, USA) equipped with a flame ionisation CH₄ detector from the Institute of Hydrobiology, Chinese Academy of Sciences. The fluxes were calculated using linear regression based on the concentration change as a function of time. 91% of all fluxes had a r-squared value of 0.70 or above (of which, 53% had a r-squared value of 0.9 or above). CH₄ flux at each site was calculated using the following equation:

$$F=\frac{d_c}{d_t}\times \frac{M}{V_0}\times \frac{T_0}{T}\times \frac{V}{A}\times 60$$

(1)

F, flux at the time of chamber closure (mg m⁻² h⁻¹);

d_c/d_t, time derivative (slope) CH₄ concentration change over time (ppm min⁻¹);

M, molecular mass of CH₄ (g mol⁻¹);

V₀, ideal gas mole volume (0.0224 m³ mol⁻¹);

T₀, absolute temperature (273.15 K);

T, absolute temperature inside of chamber at sampling (K);

V, chamber volume (m³) above the water surface;

A, chamber area (m²).

2.3. Measurement of Environmental Factors

The water depth at each site was measured with a ruler and bamboo during sampling. The air temperature was measured using a digital thermometer (TM-902C, Factory of Lihuajin Instrument, Guangzhou, China). Conductivity, water temperature, pH and DO concentration at a water depth of 10 cm were measured using a portable multi-parameter water quality meter (Multi 3630 IDS, WTW Co., Munich, Germany).

At each site, the plants were sampled in September 2021 to measure biomass. Three 50 cm × 50 cm plots were randomly selected for these measurements. The plant samples were oven-dried at 70 °C for 48 h, and then weighed. In addition, three soil samples at each site were collected at a depth of 10 cm, and all soil samples were transferred to the laboratory, air-dried indoors and then dried at 70 °C for 48 h. The samples were milled and passed through a 0.125 mm sieve to determine the organic carbon concentration (SOC, g kg⁻¹) using the wet oxidation method with K₂CrO₇, and the soil pH was measured using the potentiometric method. In addition, total nitrogen (TN) concentration (g kg⁻¹) was measured using the Kjeldahl method with H₂SO₄ digestion. The total phosphorus (TP) concentration of the soil was determined by colorimetry by alkali fusion with NaOH.

2.4. Statistical Analysis

The significant differences in CH₄ fluxes at the four sites were analysed using SPSS software (18.0 version) based on Bonferroni’s test as obtained by one-way analysis of variance. The relationship between CH₄ fluxes and environmental variables was determined using Pearson’s rank correlation. p < 0.05 was considered statistically significant.

3. Results

3.1. Environmental Factors

The mean air temperature at the ECs, OPs and ZLs sites from April to December was not significantly different (p > 0.05), with mean values of 26.5 °C, 25.5 °C and 26.9 °C, respectively, whereas the mean air temperature at the ECs and ZLs sites was significantly higher (p < 0.05) than that of the TNs site (mean value of 24.7 °C, p < 0.05) (Figure 3a). The mean water temperature at the ECs, OPs and ZLs sites was not significantly different (p > 0.05), with mean values of 23.5 °C, 23.2 °C and 23.2 °C, respectively, but it was higher (p < 0.05) than that at the TNs site (23.0 °C) (Figure 3b). The mean water depth at the ECs, OPs, TNs and ZLs sites was 151, 192, 141 and 97.8 cm, respectively, and significant differences were observed (p < 0.05); however, there were no significant differences (p > 0.05) between the ECs and TNs sites (Figure 3c). The mean pH of the water at the ECs site (7.9) was significantly lower (p < 0.05) than that at the OPs and TNs sites (8.5 and 8.2, respectively), and higher (p < 0.05) compared with that at the ZLs site (7.5) (Figure 3d). The mean DO at the ECs site (7.0 mg L⁻¹) was significantly lower (p < 0.05) than that at the Ops site (8.6 mg L⁻¹), but significantly higher (p < 0.05) than that of the ZLs site (3.2 mg·L⁻¹). There were no significant differences (p > 0.05) between the ECs and TNs sites (mean value of 7.6 mg L⁻¹) (Figure 3e). The mean conductivity of water at the ECs site (352.5 μS cm⁻¹) was significantly higher (p < 0.05) than that at the OPs (339.5 μS cm⁻¹) and TNs sites (347.6 μS cm⁻¹); however, there were no significant differences (p > 0.05) between the ECs and ZLs sites (355.5 μS cm⁻¹) (Figure 3f).

The vegetation biomass at the ECs site was significantly higher (p < 0.05) than that at the Ops site, but lower than that at the ZLs site, and higher than that at the TNs site, whereas there were no significant differences among the ECs, TNs and ZLs sites (p > 0.05) (

Table 1. The physicochemical characteristics at the four sites.

Sites	Vegetation		Soil
	Types	Biomass (g m−2)	pH	SOC/g kg−1	TN/g kg−1	C:N Ratio	TP/g kg−1
Open water (OPs)	—	No grown vegetations	8.12 ± 0.05a	16.63 ± 1.54a	1.33 ± 0.14a	12.54 ± 0.21a	0.64 ± 0.01a
Invasive plant (ECs)	E. crassipes	270.02 ± 20.64a	7.96 ± 0.04ab	29.10 ± 1.71a	2.42 ± 0.05a	12.01 ± 0.65a	0.63 ± 0.01a
Floating plant (TNs)	T. natans	211.08 ± 17.63a	7.57 ± 0.29ab	46.47 ± 11.31ab	3.44 ± 0.78ab	13.37 ± 0.49a	0.61 ± 0.01a
Emergent aquatic plant (ZLs)	Z. latifolia, N. nucifera, T. natans	618.30 ± 187.50a	7.03 ± 0.07b	56.63 ± 2.84b	4.49 ± 0.30b	12.61 ± 0.17a	0.67 ± 0.01a

Note: different lowercase letters indicate a significant difference exists among the three sites. Significance level: 0.05.

). The soil pH in the ZLs site was significantly lower than that of the Ops site (p < 0.05), and no significant differences were observed for the others (p > 0.05). The SOC and TP concentration in the top 10 cm of the soil at the ZLs were significantly higher than those at the OPs and ECs (p < 0.05); however, there were no significant differences between the ZLs and TNs sites (p > 0.05). Finally, the carbon-to-nitrogen ratio and the TP concentration were not significantly different between the sites (p > 0.05) (Table 1).

3.2. CH₄ Emission Fluxes

Temporal variations in CH₄ emission fluxes were recorded at the four sites, and the peak values occurred in the spring and/or summer, whereas the lowest values occurred in winter. The CH₄ emission fluxes at the ECs, OPs, TNs and ZLs sites ranged from 0.15 to 102.68, 0.12 to 59.75, −5.38 to 43.14 and 0.06 to 90.19 mg m⁻² h⁻¹, respectively (Figure 4).

Mean CH₄ emissions at the ECs, OPs, TNs and ZLs sites from April to December were 23.16, 9.68, 7.08 and 19.48 mg m⁻² h⁻¹, respectively (Figure 5). The highest CH₄ emissions were observed at the ECs site. These differences indicate that the invasive plant, E. crassipes, exerts an enormous influence on CH₄ emissions at Hong Lake. A one-way analysis of variance showed that there was a significant difference in CH₄ emissions between the ECs and TNs sites (P = 0.03), slight differences between the ECs and OPs sites (p < 0.1) and no significant differences between the ECs and ZLs sites (p > 0.10).

3.3. Dependence of CH₄ Fluxes on Environmental Factors

CH₄ emission fluxes at all four sites were significantly positively correlated with air and water temperature (Figure 6a, p < 0.01), and significantly negatively correlated with water depth (Figure 6b, p < 0.01). The CH₄ emission fluxes had no significant correlation with the DO concentration in the water (Figure 7, p > 0.05).

4. Discussion

In this study, CH₄ emission fluxes at Hong Lake ranged from −5.38 to 102.68 mg m⁻² h⁻¹ (Figure 4). The results indicate that this shallow lake releases a large amount of CH₄ into the atmosphere, but the flux values matched the ranges recorded in other lakes in tropical and subtropical regions (ranging from −1.7 to 326 mg m⁻² h⁻¹) [27,30]. Moreover, Gondwe and Masamba [31] reported that the maximum diffusive CH₄ emission rates in tropical wetlands varied between 0.24 and 293 mg m⁻² h⁻¹, with a mean of 23.2 mg m⁻² h⁻¹. The CH₄ flux at Hong Lake was one to two orders of magnitude higher than that at Dong Lake in the Yangtze River in China (ranging from 0.06 to 5.53 mg m⁻² h⁻¹) [32], and higher than that at lakes in North America (ranging from 0.002 to 0.826 mg m⁻² h⁻¹) [29,33] and a reservoir in the boreal region (ranging from 18 to 36 mg m⁻² h⁻¹) [34]. A hydro-electric reservoir in French Guyana released CH₄ at flux intensities ranging from 8 to 35 mg m⁻² h⁻¹ [35].

Over the last five decades, Hong Lake has been extensively altered by flood regulation, irrigation, fish aquaculture and water supply demands [23]. This has resulted in a deterioration of water quality and caused the TP and TN to exceed their protective targets (TP ≤ 0.05 mg L⁻¹ and TN ≤ 1.0 mg L⁻¹) [21]. According to a report in 2004, approximately 80% (250 km²) of the lake has been used for large-scale aquaculture since the 1990s [36]. In 2017, the area used for aquaculture decreased to 40% of the lake area as a consequence of wetland protection and a restoration project at Hong Lake beginning in 2004 [21,23]. Based on our measurements, the TP at the lake was 0.088 mg L⁻¹ in August and 0.239 mg L⁻¹ in November 2021, which could have altered the CH₄ emissions of the lake [3]. Many studies have found that eutrophic lakes release more CH₄ emissions into the atmosphere [14,37,38]. Zhou et al. [37] reported that emission values ranged from 0.1 to 351.9 mg h⁻² h⁻¹ for shallow lakes in the Yangtze River Basin and were related to more enriched waters. Pickard et al. [12] also reported that severe pollution in urban lakes resulted in higher CH₄ emissions, including a large amount of untreated sewage input, with the highest flux recorded at 335 mg m⁻² h⁻¹.

We consulted internal data from the Administration of the Hong Lake National Nature Reserve, beginning with the water hyacinth weed invasion in the 1990s. The coverage of the invading hyacinth increased from approximately 1300 ha in 2009 to 6000 ha in 2014, but decreased to 1200 ha in 2017 because of a project to remove water hyacinth from the lake. To date, it covers approximately 300 ha of the lake (Figure 8). The invasive water hyacinth weed exhibited high growth rates and a mean net CO₂ exchange (−3.4 to −5.4 g C-CO₂ m⁻² d⁻¹, negative values indicate ecosystem CO₂ uptake) compared with open water (2.3 to 5.1 g C-CO₂ m⁻² d⁻¹, positive values indicate ecosystem efflux) [39]. However, the extensive water hyacinth coverage of the lake surface resulted in the eutrophication of the lake [40]. Furthermore, the large water hyacinth mats prevented the transfer of oxygen from the atmosphere to the water and decreased oxygen production by other plants and algae [10,15]. This changes other ecological factors, such as the DO concentration in water and sediment temperature of the lake [19,20], which may be linked to CH₄ production and emissions from the lake [10]. In the present study, CH₄ emission fluxes at the ECs site were 139–227% higher on average than those at the OPs and TNs sites, and were approximately 20% higher than that at the ZLs site, although there was no significant difference between the ECs and ZLs sites (Figure 5). Our results are consistent with those of previous studies [27]. For the ECs site, higher CH₄ release rates during sampling may be explained by higher biomass. Wang et al. [30] found that the vegetation in the inundated area played an important role in CH₄ production and represented ”hotspots” of CH₄ fluxes in water systems. Table 1 shows that the plant biomass in the ECs and the ZLs sites was higher than that in the OPs and TNs sites, and the former two sites had increased CH₄ release into the atmosphere. In general, numerous studies indicate that vegetation is a key factor of CH₄ release in wetlands and is attributed to primary production, which supplies organic matter to the sediment and induces the production of CH₄ by methanogenic bacteria [41,42]. For example, Furlanetto et al. [43] reported that higher CH₄ emissions in eutrophic lakes were attributed to higher organic matter concentration, resulting from higher primary production rates. Other studies indicated that CH₄ emissions were positively correlated with net primary production in two lakes [3,32]. Our results reveal that the invasive plants strongly enhance freshwater CH₄ emissions through an increase in plant productivity, thus contributing to global warming. Therefore, to reduce CH₄ emissions, we suggest the extensive removal of water hyacinth in lakes through wetland protection and restoration projects.

CH₄ is produced in sediment under anoxic conditions by methanogens and is released into the atmosphere through three pathways, including ebullition, diffusion and plant-mediated emissions [3,44,45]. Two studies showed that lower DO concentrations in water overlaying the sediment in the lake and marsh resulted in higher CH₄ production [46,47] and greater CH₄ emissions [10,48]. In the present study, the mean DO concentration at the four sites of the lake, from highest to lowest, was as follows: OPs > TNs > ECs > ZLs (Figure 3e). Thus, lower DO concentrations were observed in surface water with areas covered by macrophytes, which may have caused an increase in CH₄ emission fluxes at the ZLs and the ECs, compared with the TNs and Ops (Figure 4 and Figure 5). The results indicate that CH₄ emissions had no significant correlation with DO concentration in the lake water, but the emissions decreased with increasing DO concentration in the lake water (Figure 7). For example, Bolpagni et al. [49] reported that the oxygen saturation in a stand of T. natans was lower than that in control areas that were devoid of plants, indicating vegetation could lead to a reduction in anoxic conditions and an increase in CH₄ emissions, although the aerenchyma in the plants may contribute to the transport of CH₄ from the water column into the atmosphere. In contrast, Attermeyer et al. [10] reported that the lower DO concentrations in the surface water of areas covered by E. crassipes led to a significantly lower flux of CH₄ emissions compared with that of open water, because the CH₄ in the surface water beneath the vegetation was oxidised by methanotrophic bacteria. Thus, the effect of DO concentrations in water on CH₄ emissions requires further study.

Air and water temperatures are important factors that limit the seasonal variation in CH₄ emissions [6,8,32,50]. It is well-established that methanogenesis in lake sediments increases exponentially with temperature because of the increased microbial activity at higher temperatures [10,51,52]. Figure 4 shows that the highest CH₄ emission fluxes occurred in June and July (Summer) at the three sites (ECs, OPs and TNs), whereas the lowest were seen in December (Winter). Figure 6a indicates that CH₄ emissions increase exponentially with air and water temperatures in Hong Lake (N = 40, p < 0.01), but R² was much lower. Although the highest CH₄ emissions occurred in April (Spring) for the ZLs site, the results suggest that other factors affect the CH₄ emissions at Hong Lake, such as water depth, plant growth and the availability of organic matter in sediments [6,45,53].

Water depth or the water table level in wetlands is usually a major factor affecting the spatial and temporal variation in the CH₄ emission flux [42,48,53]. We observed a negative correlation between CH₄ emission fluxes and water depth at all sites; however, R² was also much lower (Figure 6b). The results are consistent with the findings of several previous studies [8,53]. Because Hong Lake is a shallow lake with a mean water depth of 1.5 m [28], the variation in water depth in the lake from April to December ranged from 48 cm to 293 cm (Figure 3c). However, emergent plants, including Z. latifolia and N. nucifera plants, which were more dominant in the ZLs site compared with the other three sites, could extend their root systems into deeper and more anaerobic sediment and transport CH₄-rich gas into the atmosphere [53].

5. Conclusions

In this study, we determined how the invasive water hyacinth weed affects CH₄ emissions in subtropical Chinese lakes. The results indicate that CH₄ emissions in the zones invaded by water hyacinth were 20%–220% higher than those in the exotic-plant- free areas, because of the higher productivity of the invasive water hyacinth. The CH₄ emission fluxes at Hong Lake ranged from −5.38 mg m⁻² h⁻¹ (a sink of atmospheric CH₄) to 102.68 mg m⁻² h⁻¹ (a larger source of atmospheric CH₄). In addition, the results indicate that CH₄ emissions exhibited a weaker correlation with water temperature, water depth and dissolved oxygen concentration. Therefore, it is essential to further intensively study the CH₄ emissions of lakes in tropical and subtropical regions.