Assessment of Biophysical Properties of Faecal Pellets from Channel Catfish (Ictalurus punctatus) and Bighead Carp (Aristichthys nobilis)

Abstract

Fish faeces are a crucial component of solid wastes from cage culture systems. In order to investigate the environmental impacts of faeces from channel catfish (Ictalurus punctatus) and bighead carp (Aristichthys nobilis), certain biophysical characteristics during faecal sinking at three temperatures (10, 20 and 30 °C for winter, spring-autumn and summer conditions, respectively) were assessed in the present study. Settling velocities of faeces from channel catfish (1.72–13.33 cm/s) and bighead carp (4.16–13.83 cm/s) accelerated with an increase in water temperature. For channel catfish faeces, there were positive correlations between settling velocity and physical properties, i.e., weight, volume, length and diameter; however, for bighead carp faeces, no linear relationship between settling velocity and length was found. The main faecal water absorption period for these two species occurred after 2.5 min of immersion. The main leaching period of faecal carbon and nitrogen was 0–2.5 min, and the leaching period of faecal phosphorus was 0–10 min. The nutrient contents in channel catfish faeces were significantly higher than those in bighead carp faeces. These results suggest that co-culturing channel catfish with bighead carp can effectively reduce the discharge of nutrients from aquaculture. The biophysical properties of these two types of fish faeces can also provide guidance in particle waste collection.

1. Introduction

In China, the most common aquaculture practice in reservoirs is net cage culture, which is considered a recently developed culture system triggered by the extensive dam building that has occurred since 1949. However, cage aquaculture has been criticized for its perceived adverse environmental effects [1,2] resulting from low feed digestibility and high feed loss [3]. The release of soluble and particulate wastes derived from fish farms into the surrounding environment could cause non-point source pollution [4,5,6]. Particulate materials, mainly composed of uneaten feed and fish faeces [7], may increase nutrient levels and change the benthic ecosystem after leaching and sedimentation in reservoirs [8,9].

One approach to reduce the release of particulate wastes from the aquaculture industry is to introduce integrated multi-trophic aquaculture (IMTA) technique, which combines fed species with extractive species in a serial connection [10,11,12,13]. In China, IMTA in freshwater cages is practiced as multi-layered cage culture [14]. Channel catfish (Ictalurus punctatus) are mostly cultured in earth ponds in the USA, whereas in Mexico this species is intensively cultured in floating cages [15]. In different channel catfish aquaculture practices, the mixed-cohort culture system is more cost-effective, and results in better production than the single-cohort system [16]. As an exotic species from America, channel catfish have also been introduced to China, and have been almost exclusively stocked in multi-layered cages as fed species in many reservoirs to meet increasing domestic and export market demand. By 2020, production of channel catfish reached more than 208,000 t [17]. Bighead carp (Aristichthys nobilis) is a filter-feeding species that has been introduced into more than 72 countries [18]. Within China, the species has been successfully translocated into most provinces, with production reaching 3.13 Mt in 2020 [17]. It is usually chosen as an extractive species co-cultured with intensively fed species, in order to reduce feed loss and improve nutrient use efficiency [14]. In the USA, however, bighead carp escaped into the wild and became an injurious species, and there are no bighead carp in private aquaculture facilities anywhere [19]. In the Netherlands, this species has not been cultured or stocked on a large scale [18]. Multi-layered cage aquaculture that integrates channel catfish (feeding species) in an inner cage with bighead carp in the outer cage (non-feeding species) is gaining popularity in China [14]. Although this integration has reduced feed wastes, nutrient discharge from fish faeces still comprises an important pollution source that has usually been ignored in aquaculture deposition models [20].

Fish faeces are the main means for phosphorus outflow [21], and are the major solid waste that originates from aquaculture [22]. Early studies in nutrient discharge reported that about 15% nitrogen and 70% phosphorus were lost through fish faeces [23]. With the adoption of modern technology assisted by improved management, nutrient loss through faeces appears to be reduced [24,25]. The decreased faecal dispersion can contribute to settlement reduction, and thus become a smaller potential hazard on benthic organisms [26]. Generally, both faecal leaching and deposition processes are affected by faecal stability, which in turn is affected by physical characteristics [24] and nutrient composition. Many studies have investigated the settling and leaching characteristics of faeces from aquatic animals [7,20,26,27,28,29].

The contribution of bighead carp to the recycling of particulate wastes derived from channel catfish farming is of concern in the context of an ecosystem approach to aquaculture. During the aquaculture cycle, bighead carp mainly feed on the uneaten feed and faeces derived from the inner caged fish. However, the faeces from bighead carp also contribute to the wastes released from cage aquaculture. Due to the prevalence of these two species’ use in multi-layered cage aquaculture practices in China, we hypothesized that the nutrient contents from wastes would be much lower than those from channel catfish single-cage culture systems. Thus, in the present study, the biophysical properties and nutrient composition after immersion of these two fish species’ faecal pellets were evaluated and compared. Also, the leaching rates of key nutrients, i.e., carbon, nitrogen and phosphorus, were quantified in order to assess the environmental impact of the multi-layered cage aquaculture approach.

2. Materials and Methods

2.1. Experimental Design and Faeces Sample Collection

This study was conducted during July and December 2017 in a cage culture base (N 30°23′44″, E 111°6′30″) at the Institute of Hydrobiology, Chinese Academy of Science, which was located in the southeast of the Geheyan Reservoir in the middle of the Qingjiang River. In the laboratory, three experimental temperatures (10 °C, 20 °C and 30 °C for winter, spring-autumn and summer conditions, respectively) were chosen to reflect seasonal variances. The three temperature regimes were determined according to the annual water temperature monitoring using a temperature/light Data Logger (UA-002-64; Onset Computer Corp., Bourne, MA, USA) in the Geheyan Reservoir.

The two target species were the channel catfish (Ictalurus punctatus), which was reared in almost all farms, and the bighead carp (Aristichthys nobilis), which was cultured in the outer cages of these farms in order to reuse the uneaten feed. Faecal materials were collected from channel catfish [n = 20 fish, average weight was 629.3 g, fed on commercial feed (Jingzhou Tongwei Feed Development Company, Jingzhou, China)] and bighead carp (n = 20 fish, average weight was 1320.8 g, fed on uneaten feed, faeces and plankton). In order to obtain as undisturbed faecal pellet samples as possible, fishes were captured in situ from fish cages, anaesthetized using a solution of 100 mg/L MS-222 for 10 min, as suggested by Liu et al. (2009) [30] and Zhang et al. (2014) [31]; they were then dissected for the distal intestine upon arrival at the laboratory. The contents of the hindgut were carefully removed and divided into pieces randomly, as described in Chen et al. (1999) [27].

In the present study, the settling and leaching characteristics of these two fish species’ faeces in different seasons were explored. Prior to the determination of the settling velocity and nutrient leaching rate, faecal samples were placed on absorbent papers for 10 s and then measured for wet weight (M, g) (electronic scale PTT-A+200; precision 1 mg, HZ Electronic LLC., Springfield, MA, USA), length (L, mm) and diameter (D, mm) (vernier caliper 171–132A; precision 0.02 mm, Guilin Guanglu Digital Measurement and Control Corp., Guilin, China). A theoretical cylinder shape was assumed for the fish faecal pellet. The volume (V, mm³) of each faecal pellet was calculated as follows:

V = π × (D/2)² × L

2.2. Settling Velocity of Faecal Pellets

After the measurement of physical characteristics, faecal pellets were gently introduced into a circular section of a Perspex settling column (l = 1.3 m, internal i.d. = 0.1 m) with reservoir water and allowed to settle using forceps. The settling velocities of fish faeces for three water temperatures were determined by timing the descending distance of 1 m between two graduations, one of which was 0.1 m below the water surface while the other which was 0.2 m above the bottom, in order to ensure that the push from forceps in addition to shear effects from the bottom were negligible. For each trial, a total number of 70 faecal pellets were observed. From these measurements, average settling velocities were calculated based on the following formula:

v = 100/t

where v is the settling velocity (cm/s) and t is the recorded descending time (s).

2.3. Water Stability and Nutrient Leaching of Faecal Pellets

Preliminary research showed that the major leaching process occurs in the first few minutes of the immersion period, and that the most impacted water depth is within 50 m of fish cages [23]; thus, we analyzed the moisture and nutrient leaching rate of faeces (channel catfish faeces: n = 16; bighead carp faeces: n = 20) at 0, 2.5, 5 and 10 min. We assumed 10 min as the maximum possible leaching time based on the settling rate of fish faeces, and on the maximum water depth or impacted depth of the cage studied. Experiments on nutrient leaching rate from faeces were simulated in 500-milliliter beakers using reservoir water. After each sampling time, the faecal pellets were recollected, placed on the absorbent papers for 10 s and measured as wet weight (M₀, g). Recollected faeces were oven dried overnight at 105 °C and remeasured as dry weight (M_t, g) for the calculation of moisture (%) using the following formula:

Moisture (%) = (M₀ – M_t) × 100/M₀

After leaching, dried faeces were ground into fines, homogenized and divided into three sub-samples for analysis of carbon and nitrogen content using an elemental analyser (Flash 2000; Thermo Fisher Scientific Corp., Waltham, MA, USA), and of phosphorus content using an Inductive plasma emission spectrometer (ICAP 7000; Thermo Fisher Scientific Corp., Waltham, MA, USA).

2.4. Statistical Analysis

Settling velocity data complied with normal distribution assumptions and were analyzed using two-way ANOVA in SPSS 20.0 in order to investigate the variation between fish species and water temperature. Moisture and nutrient content data of faeces were subjected to post hoc Tukey’s test for the analysis of variance among three experimental factors, i.e., fish species, water temperature and immersion time, using R 3.1.1 (base package). The leaching rate of each nutrient did not fulfil normal distribution assumptions and was compared with Kruskal–Wallis analysis of variance followed by Mann–Whitney tests for multiple comparison analysis. SigmaPlot 10.0 (Systat Software, Inc., Palo Alto, CA, USA) was used for plotting and assessing relationships between faecal size (wet weight, length, diameter and volume) and settling velocity with a Pearson correlation test. All data were presented as mean ± SD.

3. Results

3.1. Settling Velocity of Faecal Pellets

Settling velocity of faecal pellets from channel catfish (Ictalurus punctatus) and bighead carp (Aristichthys nobilis) ranged from 1.72 to 13.33 cm s⁻¹ and 4.16 to 13.83 cm s⁻¹, respectively. Results showed significant differences (p < 0.05) in settling velocity between faecal pellets from the two fish species (Figure 1). Regarding the channel catfish, faecal pellets settled at the water temperature of 10, 20 and 30 °C with velocities (mean ± SD) of 5.34 ± 1.20 (n = 70), 6.47 ± 1.46 (n = 70) and 6.52 ± 2.04 (n = 68) cm s⁻¹, respectively. There were significant differences in settling velocity of faeces between the different water temperature treatments except between the 20 and 30 °C treatments (p > 0.05). Concerning the bighead carp, the overall average faecal settlement velocities at 10, 20 and 30 °C were 9.14 ± 1.60 (n = 70), 8.27 ± 1.86 (n = 70) and 10.10 ± 1.24 (n = 70) cm s⁻¹, respectively. The differences were significant between velocities at different water temperatures (p < 0.05).

Figure 1. Settling velocities (mean ± SD) of faecal pellets from channel catfish (Ictalurus punctatus) and bighead carp (Aristichthys nobilis) at three water temperatures (10, 20 and 30 °C). Different lowercase letters (a, b) represent settling velocities of channel catfish faeces were significantly different at different temperatures. Different capital letters (A, B, C) represent settling velocities of bighead carp faeces were significantly different at different temperatures. * represent settling velocities of these two species were significantly different.Within the same fish species (i.e., channel catfish), there was a positive correlation between faecal pellet dimensions (weight, volume, length and diameter) and settling velocity (p < 0.01 at 10, 20 and 30 °C) (Figure 2). The regression analysis for these faeces showed that the greatest correlation coefficients were found in the 30 °C treatment while the values in the 10 and 20 °C treatments were similar and lower.

Figure 1. Settling velocities (mean ± SD) of faecal pellets from channel catfish (Ictalurus punctatus) and bighead carp (Aristichthys nobilis) at three water temperatures (10, 20 and 30 °C). Different lowercase letters (a, b) represent settling velocities of channel catfish faeces were significantly different at different temperatures. Different capital letters (A, B, C) represent settling velocities of bighead carp faeces were significantly different at different temperatures. * represent settling velocities of these two species were significantly different.Within the same fish species (i.e., channel catfish), there was a positive correlation between faecal pellet dimensions (weight, volume, length and diameter) and settling velocity (p < 0.01 at 10, 20 and 30 °C) (Figure 2). The regression analysis for these faeces showed that the greatest correlation coefficients were found in the 30 °C treatment while the values in the 10 and 20 °C treatments were similar and lower.

Different regression analysis results were observed in bighead carp. The settling velocity of faecal pellets from bighead carp was correlated with faecal weight (p < 0.01 at 10 and 20 °C, p = 0.03 at 30 °C) (Figure 3). Similarly, the settling velocity was also positively, but weakly correlated with faecal volume and diameter both at 10 and 20 °C. However, neither volume nor diameter of faeces was correlated with settling velocity (p > 0.05) when temperature increased to 30 °C. Furthermore, no significant linear relationship was found between settling velocity and faecal length at three experimental temperatures (p > 0.05).

3.2. Moisture of Immersed Faecal Pellets

Figure 4 showed that the moisture content increased with immersion time for faecal pellets from different fish species. The greatest moisture content increase in channel catfish faeces was generally apparent after 2.5 min of immersion time (p < 0.05); no significant moisture increase was observed (p > 0.05) after immersion for 2.5–10 min. The moisture of faecal pellets from bighead carp was not significantly influenced by immersion time (except the difference between the non-immersed group and the 10-minute immersion group at 30 °C). Results showed that faecal material from channel catfish and bighead carp in the non-immersed group contained on average 59.66 ± 14.00% and 57.96 ± 4.75% (mean ± SD) of moisture, respectively. For channel catfish faeces, the average moisture content at 30 °C was slightly higher than those at 10 and 20 °C, although the differences were not statistically significant (p > 0.05). No consistent variation tendency was observed in faecal moisture content from bighead carp. Faecal moisture from neither channel catfish nor bighead carp was significantly influenced by water temperature (p > 0.05).

3.3. Release of Nutrients from Faecal Pellets

The proximate composition analysis of faecal pellets from channel catfish and bighead carp subjected to different leaching periods is summarized in

Table 1. Carbon content (mg g⁻¹ dry weight; mean ± SD) and leaching rate of faecal pellets from channel catfish and bighead carp after immersion at three water temperatures (10, 20 and 30 °C) for 0, 2.5, 5 and 10 min. Mean leaching rate of carbon is specified in brackets.

Fish Species and Water Temperatures	Immersion Time
	0 min	2.5 min	5 min	10 min
Channel catfish
10 °C (n = 16)	342.27 ± 12.15	329.57 ± 19.12 (3.71%)	322.24 ± 4.42 (5.85%)	314.44 ± 7.49 (8.13%)
20 °C (n = 16)	382.78 ± 3.80	378.97 ± 5.96 (0.99%)	362.28 ± 8.84 (5.36%)	349.60 ±11.98 (8.67%)
30 °C (n = 16)	375.92 ± 6.52	368.24 ± 3.47 (2.04%)	363.28 ± 7.69 (3.36%)	330.29 ± 11.01 (12.14%)
Bighead carp
10 °C (n = 20)	63.38 ± 0.75	63.26 ± 0.78 (0.19%)	59.89 ± 2.34 (5.50%)	59.76 ± 0.83 (5.70%)
20 °C (n = 20)	65.41 ± 0.45	64.04 ± 0.67 (2.10%)	61.71 ± 1.40 (5.66%)	60.68 ± 0.18 (7.23%)
30 °C (n = 20)	58.05 ± 0.37	57.73 ± 0.39 (0.55%)	56.79 ± 0.57 (2.18%)	56.59 ± 0.85 (2.52%)

,

Table 2. Nitrogen content (mg g⁻¹ dry weight; mean ± SD) and leaching rate of faecal pellets from channel catfish and bighead carp after immersion at three water temperatures (10, 20 and 30 °C) for 0, 2.5, 5 and 10 min. Mean leaching rate of nitrogen is specified in brackets.

Fish Species and Water Temperatures	Immersion Time
	0 min	2.5 min	5 min	10 min
Channel catfish
10 °C (n = 16)	16.71 ± 0.63	16.67 ± 0.13 (0.25%)	15.87 ± 0.34 (5.01%)	15.43 ± 0.69 (7.66%)
20 °C (n = 16)	24.27 ± 0.36	23.24 ± 1.43 (4.25%)	20.58 ± 0.86 (15.20%)	19.18 ± 0.28 (20.97%)
30 °C (n = 16)	26.10 ± 1.76	21.72 ± 0.82 (16.79%)	21.67 ± 0.12 (16.98%)	17.72 ± 2.36 (32.11%)
Bighead carp
10 °C (n = 20)	7.67 ± 0.10	7.40 ± 0.11 (3.54%)	7.40 ± 0.07 (3.46%)	7.24 ± 0.06 (5.66%)
20 °C (n = 20)	7.60 ± 0.09	7.51 ± 0.08 (1.17%)	7.46 ± 0.06 (1.86%)	7.32 ± 0.03 (3.67%)
30 °C (n = 20)	7.59 ± 0.33	7.28 ± 0.04 (4.08%)	7.27 ± 0.07 (4.27%)	7.18 ± 0.06 (5.41%)

and

Table 3. Phosphorus content (mg g⁻¹ dry weight; mean ± SD) and leaching rate of faecal pellets from channel catfish and bighead carp after immersion at three water temperatures (10, 20 and 30 °C) for 0, 2.5, 5 and 10 min. Mean leaching rate of phosphorus is specified in brackets.

Fish Species and Water Temperatures	Immersion Time
	0 min	2.5 min	5 min	10 min
Channel catfish
10 °C (n = 16)	25.36 ± 0.04	24.88 ± 0.05 (1.88%)	21.86 ± 0.11 (13.79%)	17.87 ± 0.04 (29.53%)
20 °C (n = 16)	23.69 ± 0.03	21.77 ± 0.05 (8.09%)	18.51 ± 0.01 (21.87%)	15.03 ± 0.05 (36.54%)
30 °C (n = 16)	20.93 ± 0.06	18.28 ± 0.07 (12.67%)	15.80 ± 0.03 (24.53%)	14.59 ± 0.08 (30.29%)
Bighead carp
10 °C (n = 20)	6.22 ± 0.03	6.10 ± 0.01 (1.89%)	5.75 ± 0.03 (7.59%)	5.68 ± 0.04 (8.71%)
20 °C (n = 20)	5.67 ± 0.01	5.47 ± 0.03 (3.52%)	5.43 ± 0.03 (4.28%)	5.43 ± 0.00 (4.32%)
30 °C (n = 20)	5.68 ± 0.02	5.57 ± 0.02 (1.99%)	5.49 ± 0.00 (3.29%)	5.47 ± 0.03 (3.75%)

. Results revealed a higher nutrient content in faeces of channel catfish than those in bighead carp, and that the highest nutrient content was found in the non-immersed group as no leaching occurred. For channel catfish, mean carbon, nitrogen and phosphorus content of faecal pellets in control groups ranged from 342.27 to 382.78 mg g⁻¹, 16.71 to 26.10 mg g⁻¹ and 20.93 to 25.36 mg g⁻¹ dry weight, respectively. Significant differences in the content of these nutrients in faecal samples were found for seasonal change under the same immersion time (p < 0.05). All of the nutrients showed a consistent decrease in concentration with increasing leaching time. The phosphorus content decreased significantly after each immersion period (p < 0.05), while the carbon and nitrogen content in the non-immersed group was mostly significantly greater than the 10-minute immersion group (p < 0.05).

For bighead carp, averaged faecal carbon, nitrogen and phosphorus content in the non-immersed control group ranged from 58.05 to 65.41 mg g⁻¹, 7.59 to 7.67 mg g⁻¹ and 5.67 to 6.22 mg g⁻¹ dry weight, respectively, at three temperatures. There were no significant differences in either carbon or nitrogen content among seasons (p > 0.05, Table 1 and Table 2), but significant differences were observed in faecal phosphorus content (p < 0.05, Table 3), i.e., phosphorus content of faeces at 10 °C was higher than those at 20 and 30 °C. Moreover, phosphorus content in faeces decreased with the extension in immersion time, and the significant differences were mainly detected between the control group and 5-minute/10-minute immersion groups (p < 0.05).

Nutrient leaching rates of faecal matter in channel catfish and bighead carp were calculated according to the nutrient content of faeces before and after immersion in reservoir water (Table 1, Table 2 and Table 3). The leaching rates of nutrient for channel catfish faeces were always higher than those for bighead carp, showing that the release of dissolved nutrients originating from channel catfish cages are generally faster at the settling stage. For channel catfish, most pellets from the same sampling temperature showed no significant differences (p > 0.05) in carbon leaching rates between immersion periods except with the non-immersed group. A significant difference was observed in the nitrogen leaching rate between the immersion times of 2.5 min and 10 min at a temperature of 20 °C. Nevertheless, the phosphorus leaching rates were significantly different between each immersion period (p < 0.05). There was no clear influence of water temperature on faecal carbon and phosphorus leaching rates, but significant differences in nitrogen leaching rates existed for experimental temperatures (p < 0.05) except between 20 and 30 °C. In brief, a large proportion of carbon loss occurred after immersion for 0–2.5 min, and for phosphorus after immersion for 0–10 min. The nitrogen loss occurred after 0–2.5 min at 10 and 30 °C and after 0–10 min at 20 °C.

Similarly with channel catfish, no significant differences (p > 0.05) were detected in carbon and nitrogen leaching rates from bighead carp faeces between immersion times except with the control group. However, the main leaching period of faecal phosphorus occurred after 10 min of immersion. The temperature experiments showed a consistent trend of lower nutrient leaching rates connected with increasing temperature. The main loss periods were at 0–2.5 min for carbon and nitrogen, and at 0–10 min for phosphorus. In general, an increase in immersion time can contribute to nutrient discharge, while the water temperature had different influences on nutrient release from faeces of these two fish species.

4. Discussion

Fish faecal settling rates were reported to be potentially affected by species, diets, water viscosity (water temperature and salinity), fish size and collection methodologies [22]. The averaged settling velocity of faecal pellets from channel catfish ranged from 5.34 to 6.54 cm s⁻¹ in a wet weight of 0.05–0.342 g, and from bighead carp 8.27 to 10.10 cm s⁻¹ in a wet weight of 0.009–0.067 g, both showing great variability in values. A higher mean settling rate was found in bighead carp faeces than channel catfish faeces (p < 0.05). The data of channel catfish in the present study were similar with the values for Atlantic salmon, both of which were higher than values from most aquatic animals in published literature [7,20,26,27,32,33,34]. The weight of faeces may be directly responsible for the faster faecal settling velocity, since the heavier faecal pellets dropped faster than the lighter ones, as other researchers suggested [26,29]. In addition, different characteristics of digestive organs are supposed to be potentially related to the variation in faecal settling velocity from different aquacultured finfish species [20]. Different digestive systems between species may result in different levels of faecal friability. The intestinal tract and physiology of lobster species are similar [35], which can explain the phenomenon that no difference was found between faecal velocity from the same species (lobsters) of different sizes [26]. In the case of different species, the settling rates of faecal pellets from seabass and seabream were far lower than that observed for salmon faeces [7,20,27,29]. Differences in diet type and feeding habits are also likely to result in the predominance of faecal pellets from bighead carp in this study. Therefore, filter-feeding species with slender intestines (such as bighead carp) can probably contribute to denser faecal pellets and faster faecal sinking rates.

Apart from fish species, temperature is another key factor that may impact faecal settling velocity. In this study, significant differences in settling velocity of faecal pellets (with exception of channel catfish for 20 and 30 °C) were detected among temperature treatments (p < 0.05), which generally increased with rising temperature. However, in the case of seabream, water temperature showed no influence on sinking rates [29].

The correlation analysis of faecal pellets from channel catfish indicated that the measured indexes, including weight, volume, length and diameter (especially the weight), were predictors of settling velocity. Temperature strengthened the correlation when it increased to 30 °C. A previous study on cod also agreed that settling velocity of faecal pellets was positively correlated with their volume, length and diameter [33]. However, as for bighead carp, only weight was a factor that could influence faecal settling velocity all year round while volume, diameter and length had little or no impact on settling velocity. The faecal mass and density instead of length and volume from the spiny lobster were reported to be related to their settling rates [26]. In contrast, diameter and length of cod faeces were estimated to be related to settling rates in an implicated model [33]. Research on salmon showed that their faecal settling velocity was correlated with none of weight, length or diameter [7,27]. The shape of faecal particles is similar to a cylinder, but becomes variable after leaching into water [20]. Thus, we propose that faecal mass rather than morphological indexes can be used as a good predictor of settling velocity. Other researchers suggested that faecal shape changes after release into water, and tends to disaggregate into smaller particles [20,27,36] which can settle out quickly or become suspended in the water column [29]. The proportion of suspended or flocculated particles is small to negligible [20,37]. In brief, it is complicated to predict faecal velocity due to the various factors involved.

Freshly evacuated fish faecal pellets absorb water during the sinking process, leading to a relatively high moisture content [27,38]. Our results showed that faecal moisture increased significantly with increasing immersion time, and the main water absorption period was within the first 2.5 min of immersion compared with the non-immersed group. The overall mean faecal moisture content was 59.66 ± 14.00% for channel catfish and 57.96 ± 4.75% for bighead carp. The water absorption process of faecal material was not influenced by temperature. In marine fish, the faeces are close to liquid in seawater [38], indicating that the water content may be higher in marine fish faeces than in freshwater species.

As a result of the high solubility and disaggregation potential, there is an increase in surface area of faecal pellets that are in contact with water during sinking, and therefore accelerated nutrient release [28]. Consequently, these nutrient resources are utilized by phytoplankton, further contributing to a higher possibility of eutrophication [39]. There was a considerable amount of ammonium and dissolved organic carbon inputted into the pelagic system through faecal dispersion in the first few minutes [28]. In salmonid farms, the faecal output rate reached 13.7% of consumed feed [40].

In order to appropriately assess the influence of two-layered cage aquaculture in reservoirs, it is urgent to obtain accurate nutritional composition of faeces and their release rates during sinking. However, the nutrients in faecal pellets collected through our methodology may have been overestimated when compared with the faeces released by fish as a result of possible reabsorption in the distal hindgut. Some research proved the same conclusion that nutrient concentrations in stripped faeces were higher than those in egested ones [41,42]. They supported the finding that egested faecal materials should be used for leaching rate experiments while stripped faecal material could be preferable for the determination of settling velocity. However, other researchers [27] showed that no significant differences were found in carbon and nitrogen content of fish faeces from any part of the rectal section, and believed that it was appropriate to use faecal samples collected from the rectum, rather than from released samples, for the determination of nutrient leaching.

The carbon, nitrogen and phosphorus content of faecal pellets from channel catfish was much higher than that from bighead carp. This result may be caused by the higher digestibility of bighead carp, which have longer digestive tracts to make the most of limited food. In addition, as one of the food sources from the surrounding environment, plankton (mainly zooplankton) may be more digestible for bighead carp. A high proportion of uneaten feed particles are rapidly eaten by co-cultured and wild fish around the farms [28]. Hence, stocking bighead carp in the outer cage can reduce the waste of uneaten feed as well as make better use of the organisms in these circumstances. In offshore IMTA, bivalves (such as mussels and oysters) are the most commonly cultivated filter-feeders placed adjacent to fish cages [22,43,44,45,46,47]. In the case of freshwater cage farming in China, it is very common to stock silver carp and/or bighead carp in order to maintain water quality [14].

Bighead carp, in spite of their ecological and economic benefits, also release large amounts of wastes by defecation, like other extractive species and wild fishes [13,28,38]. High loss rates of total carbon and nitrogen were observed after 2.5 min immersion, with the greatest loss rates after 10 min immersion in percentages as much as 7.23% and 5.66%, respectively. These values were much lower than those found in salmon faeces (22% and 26%, respectively) [7]. Similarly to nutrient content, the leaching of faecal carbon and nitrogen from channel catfish was generally faster than that from bighead carp, achieving release rates of 12.14% and 32.11%, respectively. The speed of nitrogen leaching was mostly faster than that of carbon, indicating that the leaching of nitrogen is generally faster during the initial stage of release. This result was in accordance with Fernandez-Jover et al. (2007) [28], who suggested that the loss of ammonium was faster than DOC when faeces initially contact sea water. The carbon and nitrogen content of faeces from channel catfish decreased with increasing immersion time, yet the most significant reduction was observed only after immersing for 10 min. However, some researchers indicated that the leaching of compounds from faeces may be far lower than estimated due to the short experimental duration that may limit the extent of the leaching process [28]. In our study, although faecal nutrient leaching may continue for hours after defecation [21,28,29,33,42,48], considering the settling velocity and water depth, immersion data over 10 min were sufficient for faecal waste dispersion modelling [7].

Increasing water temperature contributed to a higher nutrient content in channel catfish faeces, and thus a higher nutrient release rate; however, an opposite trend was shown in the nutrient content and release rate of bighead carp faeces. A reason for this may be that feed loss from the inner cage may be lower during warm seasons, leading to less food supply for bighead carp, and consequently a higher digestibility. Higher temperatures can promote metabolic processes, leading to higher rates of nutrient release [21]. However, no consistent pattern of faecal nutrient leaching rates from four farm-associated wild fish species were obtained at different temperatures [28]. These authors claimed that faecal pellet composition (determined by food and digestive system differences) and structure, rather than temperature, were more important factors related to nutrient leaching. Faeces with a compact structure could remain stable for a long period, thus leading to lower nutrient leaching rates.

The greatest phosphorus leaching velocity of faeces from channel catfish was 36.54%, significantly higher than that for bighead carp (8.71%). This phenomenon may be caused by the same reasons as we previously stated for carbon and nitrogen. While the main emission period of total phosphorus in bighead carp faeces was after 5 min of immersion, significant increases in phosphorus loss in channel catfish faeces were found between the immersion times of 2.5 min and 10 min. In addition, the release rate of phosphorus was greater than that of carbon and nitrogen. The main loss pathway of phosphorus was in water-soluble form (orthophosphate) within 3 h, and in Fe-, Al- and Ca-bound compounds afterwards [48]. Emission of orthophosphate was rapid in the first few days, and subsequently decreased as a result of activated bacterium [21,49]. It was reported that the loss rates of nitrogen and phosphorus in salmon faeces were 2.3% and 4.0%, respectively, and the losses of nitrogen and phosphorus through noble crayfish faeces were 4.1% and 3.5%, respectively [50]. A large amount of nutrients could be maintained in an unstable form from both fish feed and faecal sources. The major carbon and nitrogen sources were fish feed [21], while phosphorus originated mainly from faeces [21,48]. The ratios of nitrogen and phosphorus content in faeces from channel catfish and bighead carp were 1.03:1 and 1.31:1, respectively, almost five times lower than the values found in channel catfish feed pellets (5.28:1). The differences may confirm that faeces are the major source of phosphorus, but not nitrogen, and can help to determine the contribution of faeces in solid wastes.

5. Conclusions

Channel catfish grow relatively fast, while converting a significant fraction of faecal wastes into the surrounding environment. Co-culturing bighead carp in the outer cage with channel catfish in the inner cages can reduce a large fraction of uneaten feed, faeces and other resources into the surrounding water. This co-culturing practice has prevailed because of its sustainability and environmental integrity properties. In our study, the carbon, nitrogen and phosphorus contents of bighead carp faecal pellets were significantly lower than those in channel catfish faeces, which confirmed that co-culturing channel catfish with bighead carp can effectively reduce the discharge of nutrients from aquaculture. The faecal pellets’ settling velocities from bighead carp were significantly higher than those from channel catfish, while their nutrient content was lower. Thus, waste collection devices should be installed in the bottom of the outer cage. Faecal weight can be used as a predictor for faecal settling rates in these two species. Positive effects of temperature were found in the properties of channel catfish faeces (i.e., settling velocity, nutrient content and nutrient leaching rates), while the different influences were observed on the properties of bighead carp faeces (positive on settling velocity and negative on nutrient content and leaching rates). Faeces mainly absorb water and leach carbon and nitrogen within the first 2.5 min after release, while leaching phosphorus after immersion for 0–2.5 min, which indicates that fish faeces is the main origin of phosphorus wastes. Faeces also contribute a large proportion of carbon and nitrogen wastes to the environment. Therefore, considering the settling velocities of the faeces from these two species, the height of the channel catfish and bighead carp cages should not be designed and manufactured for more than 9.2 m and 13.8 m, respectively.