Water transfer projects are effective measures to solve the uneven distribution of water resources in China. However, these projects create highways for Golden Mussels (Limnoperna fortunei), which are aquatic invasive species, to invade new habitats. This has caused a series of problems such as structure corrosion, enhanced flow resistance, pipe blocking, and water pollution. In order to design remediation strategies, it is important to understand the factors influencing the Golden Mussel invasion. This paper investigates the invasion of Golden Mussels in the East River Water Source Project and the Northern Inter-basin Water Transfer project in China. The maximum densities of Golden Mussel attachment in these projects are about 10 000 ind. m−2. Flow velocity was found to be a significant influencing factor: the optimal flow velocity for attachment is in the range of 0.3 to 0.9 m s−1. Water temperature was also an important factor: the water temperature in the cold season influences the density of Golden Mussels and additionally, their growth and shell size are related to food competition between different individuals and water temperature. There was no obvious relationship between the pressure and the attachment density of Golden Mussels. Because all data was obtained during an infrequent pause in the water transfer projects, the data of this study is limited and the conclusions are preliminary descriptive observations. For future management decisions, more rigorous studies should be conducted to develop quantitative models aimed at providing a confident basis for controlling Golden Mussel invasion in water transfer projects.
Aquatic ecosystems are important vectors of material and energy fluxes and are essential in maintaining the biodiversity and ecological balance. Along with the intensification of human activities, aquatic ecosystems have been highly disturbed by the introduction and spread of non-native invasive species (Nakano and Strayer, 2014). Many aquatic invasive species have had damaging effects on human interests (Havel et al., 2015).
The Golden Mussel (Limnoperna fortunei, Dunker, 1856) is an aquatic invasive bivalve that is characterized by fast growth and strong reproductive ability. It possesses very broad environmental tolerance and can survive in environments with water temperature from 0 to 35 °C, flow velocity from 0.1 to 2 m s−1, water depth from 0.1 to 40 m, dissolved oxygen from 0.2 to 11.33 mg L−1, and limited food availability (Darrigran et al., 2012; Xu et al., 2015). The invasion of Golden Mussels disturbs the ecological balance of the invaded water by perturbing the local food web (Nakano et al., 2015), modifying the nature and complexity of the substrate (Borthagaray and Carranza, 2007), and then influencing the macroinvertebrate structure (Darrigran and Damborenea, 2011).
Water transfer projects are widely used to solve problems of uneven water distribution. However, water transfer projects create highways for the invasion of Golden Mussels into new habitats, resulting in their rapid spread (Xu et al., 2015). The Golden Mussel larvae enter into water transfer projects with water flow in the breeding stage, and gradually attach to the structure (typically the walls, gates, etc.) after the plantigrade veliger stage. Thanks to its byssal threads, this species is able to aggregate at high densities to the structure (Boltovskoy, 2017), thereby causing important adverse effects on the operation of water transfer projects. First, the mussel attachment increases flow resistance and can ultimately block the cross-section of the flow. The water-cooling pipes of Wuhan Iron and Steel Company (China), for example, were blocked in the 1970s by the attachment of Golden Mussel (Group of Pipeline Study, 1973). Second, the mussel attachment corrodes the concrete walls of water channels and weakens their structural strength (Perez et al., 2003). Third, the Golden Mussels pollute the water by their respiration and decay after death (Darrigran, 2002).
An adequate choice of the operation mode and parameters (velocity, temperature, pressure, light, etc.) of water transfer projects may be an efficient strategy to limit the invasion of Golden Mussels. Therefore, it is important to identify the main factors influencing their invasion which has been investigated in foregoing research. Physical and chemical parameters such as water temperature, water velocity, total nitrogen, electrical conductivity, phosphate, and dissolved oxygen were found to influence the attachment density and growth of Golden Mussels (Darrigran et al., 2012; Ernandes-Silva et al., 2017). Hot water (42-51 °C) kills mussels (Perepelizin and Boltovskoy, 2011), and suspended inorganic sediment decreases the quality of suspended food and hampers their expansion (Tokumon et al., 2016). Biological factors also affect the Golden Mussels; some fish species, for example, feed on them and thereby reduce their density. Foregoing research has focused only on water transfer projects with a length of less than 200 km (in one basin) (Nakano et al., 2012; Xu, 2012). At present, there are no studies involving a length longer than 1 000 km (inter-basin). Therefore, considering there are different scales of water transfer projects in China, one small-scale (in a basin) and one large-scale (inter-basin) project were chosen to conduct systematic research on Golden Mussel invasion.
The objectives of this study are: 1)) to analyze the influence of flow velocity, pressure and light on the attachment density of Golden Mussels; 2) to reveal the influence of water temperature on the attachment of Golden Mussels; 3) to provide suggestions for improved project management and operation.
Study area and methods
East River Water Source Project
The East River Water Source Project (ERWSP) in Southern China transfers water from the East River and its tributary the Xizhijiang River to Shenzhen City (Fig.1). Table 1 lists the basic parameters of the ERWSP. The ERWSP is situated in a subtropical maritime monsoon climate zone. The annual average precipitation is 1 500-2 400 mm, and the annual average temperature is 21-24 °C. The total length of this project is 107 km, including 17 km double steel pipes, an 80 km long concrete tunnel, several sections of culverts, and numerous valves. Most of the ERWSP consists of pipe-flow reaches where the flow is driven by pressure gradients, but there are also open-channel reaches where the flow is driven by gravity. The flow velocity along the project varies in the range 0.02-1.59 m s−1. This project has suffered from the Golden Mussel invasion since it started operation in 2001.
Northern inter-basin water transfer project
The northern inter-basin water transfer project (NIWTP) transfers water from central China to northern China (Fig.1). Table 1 lists the basic parameters of NIWTP which runs through the subtropical monsoon climate and the temperate continental climate zones. The climate difference along the project is significant with different annual average precipitation and annual average temperature. This project has a total length of 1 432 kilometers, and runs through the provinces Henan, Hebei, Beijing and Tianjin. The canal is high in the south and low in the north, allowing for gravitational open-channel flow over most of the pathway. Flow velocities along the project are in the range 0.33-1.1 m s−1. This project has suffered from the Golden Mussel invasion since it started operation in 2014.
Golden Mussel adults were sampled from 54 representative sections along the ERWSP from October to November 2009. These sections were located at distances from the East River of 0-107 km. Three 0.10 m × 0.10 m sample squares were selected for each section, and the number of adults was counted. The attachment density in each cross-section was obtained as the average density of the three samples.
Golden Mussel adults in the NIWTP were sampled from October to December 2017 in 12 cross-sections. Their distance from the canal head is indicated in Table 2. Considering the long distance of the NIWTP, the attachment density in each cross-section was obtained as the average density of nine samples in 0.1 m × 0.1 m squares. Three squares were located in the cross-section itself, three in a cross-section 400 m upstream, and three in a cross-section 400 m downstream. Highly representative cross-sections located at 1 199.4, 1 276.9 km were not accessible and could not directly be observed and sampled. In these cross-sections, a Sirio underwater robot (ROV, Di Risio et al., 2018) was used to conduct auxiliary observation with a high-speed underwater camera system. The attachment density in these cross-sections was estimated from the images.
For the sections with pipe-flow, the pressure was determined from the project design data. The average flow velocity of each section was obtained by dividing the flow discharge by the cross-sectional area. Water transparency was determined using a Secchi Disc. The shell length of the Golden Mussels along the NIWTP was measured by a Vernier Caliper. For the inaccessible cross-section, it was estimated from the images taken.
The ERWSP is appropriate for investigating the influence of velocity on the attachment density, because significant variations in velocity occur along the pathway and water temperature is similar. The ERWSP contains pipe-flow reaches with significant variations in pressure from 1 to about 20 mwc (meter of water column). Along the pathway and open-channel reaches subject to atmospheric pressure (zero relative pressure) (Table 1). This makes the ERWSP appropriate for investigating the influence of pressure. Because pipe-flow reaches are typically very dark, whereas open-channel reaches are exposed to daylight, the ERWSP is also appropriate to investigate the influence of light.
The NIWTP is an inter-basin water transfer project, that is particularly appropriate for investigating the competing influences of water temperature on the Golden Mussel attachment density and shell size.
All statistical analyses were performed using Microsoft Excel 2016 software (Microsoft, USA) except shell length analysis. Shell length distribution was analyzed using origin 8.0 software (Origin Lab, USA).
Results and discussion
The attachment density along the East River Water Source Project and Northern inter-basin water transfer project
Figure 2a shows the attachment density along the ERWSP. The density near the intakes is extremely high: about 5 000 ind. m−2 near the East River intake and about 10 000 ind. m−1 near the Xizhijiang River intake. The attachment density sharply decreases with distance from the two intakes, with values below 100 ind. m−2 at a distance of 29 km. Similar observations were made by Nakano et al. (2012). Reaches of high attachment density occur between km 57-62 (500 ind. m−2), km 86-100 (3 500 ind. m−2) and km 105-107 (5 000 ind. m−2), which is just downstream of reservoirs located at km 57, km 86 and km 105. These reservoirs act as stepping stones that provide suitable habitat that allows Golden Mussel larvae to maintain their population. Moreover, they provide planktonic larvae that can invade the reach downstream. This explanation of the locally increased attachment density is in line with Havel et al. (2005) and Nakato et al. (2015).
Figure 2b shows the attachment density along the NIWTP. The density is less than 2000 ind. m−2 from NI1 to NI4, increases to a maximum of 10 190 ind. m−2 at NI5, stays above 5 000 ind. m−2 onto NI8, and has negligible values from NI9 on. There are no reservoirs or other lentic environments along the NIWTP. The strong increase from NI4 to NI5 might be attributed to other conditions.
Influence of flow velocity on attachment density
Figure 3a shows the relation between the attachment density and the flow velocity along the ERWSP. The flow velocity has an influence on the attachment density of Golden Mussel. When the flow velocity is too low (<0.3 m s−1) or too high (>1.4 m s−1), the attachment density of Golden Mussel is very low. The low density at low velocities is tentatively attributed to insufficient food supply at low velocities. The low density at high velocity is tentatively attributed to the high boundary shear stresses that hamper the Golden Mussels attaching to the walls. The high attachment densities mainly occur when flow velocity was in the range of 0.3-0.9 m s−1, which is interpreted as the Golden Mussel’s suitable velocity range.
Figure 3b shows the relation between the attachment density and the flow velocity along the NIWTP. The flow velocity of all sites except NI8 (1.1 m s−1) is 0.33-0.55 m s−1, which is in the suitable range and only showed gradual variations. Therefore, the flow velocity cannot be the dominant factor of influence for changes in density, and especially for the strong density increase from NI4 to NI5 (Fig. 2b).
Influence of pressure and light on attachment density
Most of the ERWSP consists of pipe-flow reaches, with a pressure in the range of 0 to 20 mwc. Figure 4 shows the relation between attachment density and pressure in the ERWSP. In the open-channel reaches (P = 0 mwc), the attachment density ranges from 200 ind. m−2 to 1 500 ind. m−2. In the pipe-flow reaches with a pressure of 10 mwc, the attachment density ranged from 0 to 10 000 ind. m−2. In general, the attachment density seems to be higher in pipe-flow reaches than in open-channel reaches, which complies with the experience of the project manager (Li and Su, 2007). No clear relation seems to exist, however, between the pressure and the attachment density in the pipe-flow reaches. These observations suggest that the dominant influence is not exerted by pressure, but rather by light: pipe-flow reaches are typically very dark while open-channel reaches are exposed to daylight. Xu (2012) and Duchini et al. (2015) found that adult Golden Mussels prefer dark environments and actively try to migrate towards dark areas. Xu (2012) also found that larvae are unable to adapt to strong light.
Table 2 lists the water depth and water transparency at the sampling sections of NIWTP. The water transparency in open-channel reaches of the NIWTP was estimated at 1.6 ∼ 2 m, which is considerably smaller than the water depth, which ranges from 3.37 m to 5.70 m in the open-channel reaches. In the field investigations, it is found that the attachment occurs mainly at depths of 3 m or more below the water surface. It can reasonably be assumed that the light intensity at the bottom of the open-channel reaches is weak and not a dominant parameter of influence that causes changes in attachment density along the channel of NIWTP. It is worth noting that the attachment density in NI8 is still as high as in NI7 and NI6 at about 5 550 ind m−2 (Fig. 4), although the velocity of 1.1 m s−1 in NI8 is unsuitable. This can be explained as: NI8 is a dark underground pipe-flow reach.
Influences of water temperature on attachment
Figure 5 shows the water temperature in April-November and December-March at the investigated sections of the NIWTP. The water temperature in April-November is always higher than 10 °C, above which mussel survived (Oliveira et al., 2010) and have little effect on attachment density along the NIWTP. The larval density at sampling sites NI5-NI8 (369 ± 32 ind. m−3) is higher than that at NI1-NI4 (180 ± 32 ind. m−3), which could explain the strong increase in attachment density at NI5-NI8. Surprisingly, although the larval density at NI9-NI12 (1 165 ± 501 ind. m−3) is higher than NI5-NI8, the attachment density of sampling sites NI9-NI12 is the lowest. This seemingly incoherence can be explained by the water temperatures in December-March. Water temperatures at NI1 to NI8 remain higher than 5 °C and allow the planktonic larvae to settle (Mackie and Brinsmead, 2017), while they reach values significantly below 5 °C that strongly reduce the mussel density at NI9 to NI12. These observations demonstrate that the attachment density is influenced by history effect on the seasonal time scale, and not only by the instantaneous climatic conditions.
Figures 6 shows the average shell length and its distribution along the canal of NIWTP. The average shell length at NI1-NI4 is largest with 21-22 mm, followed by NI5-NI12 with 16-17 mm. Similarly, the shell length with highest frequency shows the same trend and is 20-24 mm at NI1-NI4, 14-20 mm at NI5-NI8, and 16-20 mm at NI9-NI10.
The shell size is influenced by two competing factors: water temperature and food availability. Water temperature is known to exert a dominant control on the growth rate, with higher temperatures leading to higher growth rates (Nakano et al., 2011). The decrease of the average water temperatures in December-March from the south to the north along the NIWTP favors a decrease in shell size. Beside temperature, also food availability contributes to the changes in shell size from NI5 to NI12. Due to the higher attachment density of Golden Mussel at NI5-NI8 as compared to NI9-NI12, less food is available for growth per individual. This explains why the shell size at NI9 to NI12 is similar to that at NI5-NI8, in spite of the marked decrease in water temperature in December-March.
Mitigation strategies in water transfer projects
In the last decades, a lot of research has been conducted to find strategies to mitigate the attachment and invasion of Golden Mussels all over the world. Various chemical and physical measures of avoiding Golden Mussel's attachment were taken, such as applying pesticide (Darrigran et al., 2011), spraying with hot water (Morton, 1982), removing artificially (Xu et al., 2015). In addition to the commonly used physical and chemical methods, some new prevention methods have been proposed in recent years, such as using coatings exhibiting very low preferential mussel attachment (Amini et al., 2017), applying an integrated ecological prevention method (Xu et al., 2015), killing Golden Mussel veligers by different turbulence-generating materials (Zhang et al., 2017).
This study gave preliminary descriptive observations of attachment of Golden Mussels in two water transfer projects of China. It is found that water velocity, light, water temperature, and food availability all influenced the attachment of Golden Mussel in water transfer projects. Providing unfavorable conditions, such as abundant light, unsuitable velocities, cooling, etc., may be a good strategy for controlling Golden Mussel invasion in projects. Because all the data was obtained in the field investigations during the infrequent stop of water transfer projects, the data of this study is limited and the conclusions are preliminary. For future management decisions, more rigorous study should be conducted to develop quantitative models aimed at providing a confident basis for controlling Golden Mussel invasion in water transfer projects.
The Golden Mussel (Limnoperna fortunei) is an aquatic invasive species. Its invasion in water transfer projects has resulted in complications such as structure corrosion, enhanced energy losses and pipe blocking, and water pollution. In order to design effective mitigation strategies, it is important to study the attachment of Golden Mussels in water transport projects and its influencing factors. The present paper reported an investigation in two water transfer projects in China that are affected by the Golden Mussel invasion: the ERWSP with a length of 107 km and the NIWTP with a length of 1 432 km.
Four dominant factors of influence were identified in the present investigation: velocity, temperature, light and food availability. The flow velocity has a significant influence on the attachment density of Golden Mussels: the optimal flow velocity for attachment is 0.3-0.9 m s−1, whereas lower and higher velocities prevent attachment. Temperature exerts a crucial influence in the cold season. The water temperatures below 5 °C strongly reduces the mussel density. Pressure does not influence the Golden Mussel attachment. The higher attachment density in pipe-flow reaches than in open-channel reaches is not due to pressure difference, but rather to differences in light exposure: the Golden Mussel prefers dark environments. A reduction in food availability per individual and water temperature leads to smaller shell sizes.
This study gave preliminary descriptive observations of Golden Mussel attachment in two water transfer projects. For future management decisions, more rigorous studies should be conducted to develop quantitative models aimed at providing a confident basis for controlling Golden Mussel invasion in projects.
This study was financially supported by the State Key Laboratory of Hydroscience and Engineering Project (2019-KY-01), the National Natural Science Foundation of China (51809086, 51779120), Key Research Project of the Higher Education Institutions of Henan Province (16A416002).