Surface Water Quality in Aquacultural Areas in an Giang Province, Vietnam

This study was conducted to assess the impact of aquaculture with three fish cultivating models including net fence, cage and earth-pond to surface water quality, using water monitoring data from 2011 to 2019 provided by Department of Natural Resources and Environment, An Giang province, Vietnam. Water quality parameters included temperature, pH, dissolved oxygen (DO), total suspended solids (TSS), biochemical oxygen demand (BOD), nitrate (NO3¯-N), orthophosphate (PO43--P) and coliforms at 13 locations were evaluated followed the National Technical Regulation on surface water quality (QCVN 08-MT: 2015/BTNMT). Multivariate analysis methods comprising Cluster Analysis (CA) and Principal Component Analysis (PCA) were used to group sampling locations according to pollution levels and to identify main water variables influencing on surface water quality. In the aquacultural areas, DO was low while TSS, BOD, PO43--P and coliform were high comparing to QCVN 08-MT: 2015/BTNMT. Among the three types of aquaculture, earth-pond culture resulted in more serious environmental pollution than cage and net fence. Five sources of pollution in the studied water bodies were identified using PCA in which temperature, pH, DO, TSS, BOD, NO3¯-N, PO43--P and coliforms could reflect the quality of water environment affected by aquaculture. CA finding suggested that the number of monitoring points could be reduced from 13 to 9 sampling locations, thus reducing monitoring cost. Future studies should focus on investigating sources of surface water pollution in the aquacultural areas.


INTRODUCTION
An Giang is one of the localities highly potential for economic development in the Mekong Delta region. The dense rivers, canals and abundant fishery resources have generated favorable conditions for the province to increasingly develop its potential, especially its strengths in aquaculture. Approximate140 species of fish have been added during the annual flood season (Nhi, 2005). According to the Department of Agriculture and Rural Development of An Giang province, the area of aquaculture in 2013 reached 2,496 hectares, including the main cultured species of Tra catfish (Pangasius hypophthalmus), giant freshwater prawn (Macrobrachium rosenbergii), snakehead fish (Channa striata), basa (Pangasius bocourti), grouper (Pangasius conchophilus), he fish (Barbonymus altus), tilapia (Oreochromis niloticus). Of which, Pangasius hypophthalmus is the main fish for farming in the province with an area of 1,119 ha with a harvested fish output of 273,939 tons equal to 120% compared to 2018. In the first nine months of 2019, the socio-economic situation of the first nine months of 2019 reached 361,000 tons, increasing 4.6% (+ 15.8 thousand tons) over the same period. Of which, cultivation were 347.6 thousand tons, increased 5.8% (Pangasius hypophthalmus accounted for 287.6 thousand tons, increased 7.4%); exploitation was 13.6 thousand tons, equaling 87.8% over the same period, due to low flood water this year reducing natural aquatic resources. Currently, most aquaculture areas do not have water treatment system and cleaning the wastewater maily relies on self-purification of water.
Aquaculture is often associated with environmental pollution if it is not well managed. One of the common environmental problems is the release of excess nutrients into the environment. Thich (2008) reported that the amount of N discharged from intensive Pangasius hypophthalmus ponds was 57.3% (5.43% in water, 50.4% accumulated in sediment and 1.5% losses due to evaporation or infiltration) and the amount of P release was 70.2% (1.8% in water, 64.5% in sediment and 3.9% losses In addition to the impacts of aquaculture on surface water quality, natural activities such as erosion, stormwater runoff, production and business activities, residential and urban areas, industrial production activities also affect surface water quality (Nga, 2009;Ly and Giao, 2018;Truc, 2019).
Multivariate analysis methods including cluster analysis and principal component analysis are widely applied to assess water quality according to space and time at several monitoring locations using multiparameters (Zeinalzadeh and Rezaei, 2017). Multivariate analysis is used to assess changes in river water quality and identify sources of pollutants (Chounlamany et al., 2017). In addition, the multivariate analysis method is used to establish water monitor network and identify water quality variables that cause changes in surface water quality (Zeinalzadeh and Rezaei, 2017). This study was conducted to assess surface water quality in aquaculture areas in the form of net fence, cage and earth-pond between 2011 and 2019. The results of the study could provide scientific information on progression of water quality due to the impact of aquaculture for better management of water quality in aquacultural sector.

Water quality data
The data on surface water quality in An Giang province was collected at the Department of Natural Resources and Environment of An Giang province in the period of 9 years from 2011 to 2019 at 13 locations in three types of aquacultural farming including three locations in the net fence model (TS7, TS8, TS9), three locations in the cage model (TS1, TS3, TS4) and seven locations in the earth-pond model (TS2, TS5, TS6, TS10, TS11, TS12, TS13). The sampling locations were shown in Figure 1. Water quality parameters including temperature, pH and dissolved oxygen (DO) were measured directly on site by handheld device. The water quality parameters such as biological oxygen demand (BOD), nitrogen nitrate (NO3 --N), orthophosphate (PO4 3--P), and coliforms were collected, properly stored and transported to the laboratory for the analysis using standard methods listed in Table 1.

Data processing
The difference in the mean value of the water quality parameters was performed by analysis of variance (ANOVA), at the significance level of 5%, Duncan test (Ahrari et al., 2015) using statistical software IBM SPSS statistics for Windows, Version 20.0 (IBM Corp., Armonk, NY, USA). The two main methods used in the evaluation of monitoring data are Cluster Analysis (CA) and Principal Component Analysis (PCA). The cluster analysis method is applied to group water locations based on all physical, chemical and biological criteria. Sampling sites with similar pollution characteristics will be grouped into the same group, while different pollution characteristics will be grouped into another group and presented as a structural tree (Feher et al., 2016;Chounlamany et al., 2017). PCA method is used to reduce initial data variables that do not make an important contribution to data fluctuation while creating a new group of variables called principal components (PC). These PCs are not related to each other and appear in descending order of importance. The important value to consider in principal component analysis is the eigenvalue coefficient, the greater the coefficient, the greater the major contribution to explaining the variability of the original dataset. This technique is used to determine the number of sources that affect surface water quality in environmental monitoring (Feher et al., 2016). The correlation between the principal component and the primary data variables (water quality indicators) is shown by the significant correlation coefficients (Feher et al., 2016). The absolute value of the significant correlation coefficient is greater than 0.75, which means that the strong correlation between the principal components and water quality criteria, from 0.75-0.5 is the average correlation, and 0.5-0.3 is the weak correlation (Liu et al., 2003). CA and PCA are implemented using copyrighted Primer Software V5.2.9 (Plymounth, UK).

Water quality in the fish cultivating areas
The surface water quality affected by cultivating models were presented in Tables 2-4. Water temperature at 13 sampling locations in three models were in the range of 29.8±1.3-29.9±1.0 o C, 29.4±0.7-29.6±1.0 o C, 29.0±0.7-30.3±1.1 o C in the cage, net fence and earth-pond models, respectively in the period of nine years. Previous study showed that temperature in Hau River ranged from 27.1-32.0 o C (Giao, 2020) and Mekong River ranged from 19.9 to 32.2 o C (Ongley, 2009). Water temperature at the study sites is still suitable for aquatic life according to National technical regulation on surface water quality (QCVN 08-MT: 2015/BTNMT) and Boyd (1998). The mean pH values in the two areas influenced by cage model (7.1±0.2-7.2±0.2) and net fence (7.0±0.2-7.2±0.3) were not much different over the years as well as the sampling locations, fluctuating around the neutral value. In the earth-pond area, the mean pH ranged from 6.7±0.3 to 7±0.3 which is slightly lower compared to the other models. Previous studies reported the pH values in Tien river and Hau river ranged from 7.1 to 7.2, 6.7 to 7.12 (Giao, 2020    ). In the current study, the average TSS concentrations in the earth-pond areas were more polluted than those at the cage and net fence areas. The findings revealed that aquacultural activities partly contributed to high TSS in the water bodies which could be artributed to the direct discharge or improperly treated of wastewater generated from aquacultural sites to the receiving waters.
The mean BOD concentrations were in the ranges of 6.7±1.5-7.6±2.5 mg/L, 7.3±3.8-10.7±6.6 mg/L, and 7.5±2.6-21.5±12.0 mg/L in the cage, net fence and earthponds models, respectively ( Table 2 -4). BOD at the water bodies influenced by earth-pond cultural practices was found higher than those in the water bodies influenced by cage and net fence models. Previous studies reported that BOD concentration in water bodies in An Giang province was in the range of 6.6 ± 1. sampling points over nine years in the current study were in the suitable ranges for aquatic life. However, high concentrations of NO3 --N can potentially cause eutrophication which subsequently lead to degrade surface water quality resulting in unfavorable conditions for aquatic organisms.
Orthophosphate (PO4 3--P) concentrations ranged from 0.1±0.1 to 0.2±0.2 mg/L, 0.1±0.0 to 0.3±0.3 mg/L, and 0.21±0.17 to 1.68±1.01 mg/L at the water bodies influenced by cage, net fence and earth-pond models, respectively (Table 2-4). Concentrations of PO4 3--P in the water bodies influenced by the earth-pond model was much higher compared to those influenced by cage and net fence models. PO4 3--P agricultural areas and in Hau River ranged from 0.02 to 0.47 mg/L (Ly and Giao, 2018), in Hau River section from An Giang to Hau Giang provinces ranged from 0.04 to 0.11 mg/L (Giao, 2020), and in canals in Soc Trang province ranged from 0.05 to 0.9 mg/L (Tuan et al., 2019). The former studies and the current study revealed that PO4 3--P at the water bodies in the Vietnamese Mekong delta exceeded the QCVN 08-MT: 2015/BTNMT, column A1. The eutrophication is very likely to take place onece the PO4 3--P concentration is greater than 0.1 mg/L (Boyd and Green, 2002). This study found that surface water quality influenced by the aquacultural activities is at risk of eutrophication due to orthophosphate concentration.

Key parameters influencing water quality at the fishery cultivating areas
A total of eight variables of water environment parameters were included in the PCA. The key water parameters influence on surface water quality were presented in Table 5. PC1 explained 58.4% of the variation in the data obtained. The source of this pollution is mainly reflected in the pH (0.420), TSS (-0.446), and BOD (-0.400). PC1 could explain for the water quality at the study site being polluted organically due to the residues of aquatic feed and untreated sewage in the areas of earthponds discharged directly into the receiving water. PC2 explained 18.7% of the fluctuation of water quality through temperature (0.50) and NO3 --N (-0.673). PC3 explained 11% of the data fluctuation, is affected by a moderate level of temperature (0.549), a weak level of DO (-0.445), and a PO4 3--P (-0,424). PC4 explained 7.4% of the data fluctuation, is affected by moderate level of temperature (0.575), DO (0.534) and weak level of NO3 --N (0.441). PC5 explained only 2.3% of the water quality variation, affected by the weak level of DO (0.464) and the high level of coliforms (0.772), and showed that microbiological pollution at the surveyed locations. The results also showed that the change of temperature is influenced by at least three factors (PC2, PC3, and PC4), DO by three factors (PC3, PC4, and PC5) and NO3 --N by two factors (PC2, PC4). The other indicators were only influenced by one PC. Thus, there are at least five sources of pollution affecting water quality in the water bodies that receive aquaculture wastewater. Water pollution sources could be hydrological regimes (currents, tides, and flows), erosion, aquaculture waste, and domestic waste. Eight monitoring indicators including temperature, pH, DO, TSS, BOD, NO3 --N, PO4 3--P and coliforms are all very important for monitoring water quality in the water bodies receiving aquacultural wastewater.

Grouping water quality in the aquaculture area
The results of water quality clustering were presented in Figure 2. At the distance of 4.0, water quality in the water bodies receiving aquaculture water can be classified into three groups. Group 1 includes the site TS5; Group 2 includes the locations of TS1, TS2, TS3, TS4, TS6, TS7, TS8, TS9; and Group 3 with the positions of TS10, TS11, TS12 and TS13. From this result, it is possible to select a water sampling location based on the locations in the same group with similar water quality, so only a representative location is needed. However, in this case, due to the locations in different water bodies such as Tien River (TS1, TS2), Chau Doc River (TS3), Hau River (TS4, TS6, TS7, TS8, TS9), Xang Vinh Tre canal (TS5), Xa Doi Canal (TS10), Don Dong canal (TS11, TS12) and Tra Cu Canal (TS13), so the locations should be arranged according to the same rivers. On Tien River, it is possible to choose TS1 or TS2 because these two locations are grouped into Group 2; On the Hau River, only one sample from TS7-TS9 is required to be sampled and one of the three positions from TS7-TS9 is located in the same subgroup in Group 2. The two positions on the Don Dong channel are in the same group with Group 3, so only one is needed. Thus, from 13 sampling points can be reduced to 9 sampling points according to this study.

IV. CONCLUSION
The analysis of surface water quality affected by aquaculture activities in An Giang province in the period of 2011 -2019 showed that the water has the problems of total suspended solids, organic matters and coliforms. Among the three types of aquaculture, earth-pond culture resulted in more serious environmental pollution than cage and net fence since the water quality assessment criteria such as DO, TSS, BOD, NO3 --N, PO4 3--P and coliforms in earth-pond culture area were much higher than those at the cage and net fence areas. PCA analysis showed that there are five main factors corresponding to five sources of pollution in the studied water bodies. The parameters of temperature, pH, DO, TSS, BOD, NO3 --N, PO4 3--P and coliforms are very important for monitoring water quality in the area receiving aquaculture wastewater. CA analysis showed that water quality in aquaculture-affected water bodies can be classified into three groups and from this grouping, the number of monitoring points can be reduced from 13 to 9 sampling points taking into account the different water bodies. Future study should focus on investigating specific sources of pollution to effectively solve the pollution problem in the study area.