The present study reveals significant temporal and spatial differences in the physicochemical and biological properties that are considered as key drivers of the marine environment. To allow for better comparison and analysis of the data, a categorization based on three-year phases (i.e. 2013-2015, 2015-2017 and 2017-2020) has been adopted. The parameter contribution was observed using a logarithmic transformation and the results are presented in Figure 2 (a,b,c).
To assess the variability of each parameter over different years, the data were analysed using boxplots as shown in Figures 3a to 3r. The box plots were labelled PR (pre-monsoon), MO (monsoon), PO (post-monsoon), and Y (year) for the respective years. The analysis showed that most parameters showed significant deviations in 2019 and 2020, with the exception of nitrite (Fig.3i), salinity (Fig.3g), faecal coliform bacteria (Fig.3l), mercury (Fig.3n) and chlorophyll (Fig.3m). This observation is of concern and suggests that the discrepancies may be due to municipal and river discharges from land into the estuarine-sea ecosystem.
On the other hand, parameters such as total suspended solids (TSS) (Fig.3e), turbidity (Fig.3f), total organic carbon (TOC) (Fig.3h), nitrite (Fig.3i) and faecal coliform (Fig.3l) showed maximum variability in 2014, 2015 and/or 2016 (Fig. 3). This variability is due to the Mahanadi River floods in 2014-2016. Furthermore, the boxplots comparing seasonal data (Figures 3a to 3r) showed that in 2014, 2015, 2016 and 2019 pH (Fig.3a) showed a typical trend of pre-monsoon < monsoon < post-monsoon while showing another trend in other years. Salinity (Fig.3g) consistently showed lower values during the monsoon season throughout the study period. Nitrite (Fig.3i) had unusually high levels during the 2015 and 2016 pre-monsoons compared to other years, which had a significant impact on the dataset. This anomaly can be attributed to uncontrolled discharges from fertilizer plants in those particular years, later regulated by the State Pollution Control Board (OSPCB). Phosphate (Fig.3j) showed higher levels during the 2019 pre-monsoon, suggesting that special attention and limitations are needed due to possible contributions from fertilizer crops.
Chlorophyll-a (Fig.3m) showed seasonal variations with a trend of pre-monsoon > monsoon > post-monsoon in 2013, 2015 and 2017, possibly due to upwelling and nutrient enrichment. However, in 2019 and 2020 the trend shifted towards pre-monsoon < monsoon < post-monsoon. The cadmium (Fig.3r) concentrations did not show any significant seasonal fluctuations, but the concentration was significantly higher in the period 2019-2020 compared to the previous years (2013-2018). Whereas, iron (Fig.3p), manganese (Fig.3o), and lead (Fig.3q) showed strong seasonal variation during 2019 & 2020, 2016 & 2020, and 2018 & 2019, respectively. However, their concentrations were significantly higher in 2019-2020 compared to 2013-2018. In contrast, mercury (Fig.3n) concentrations were minimal in 2019-2020 compared to 2013-2018. It showed strong seasonal variation during 2014, 2015, 2016, 2017, and 2018. The significant change in the chlorophyll-a trend during the pre-monsoon and post-monsoon periods in the first phase (2013-17) compared to the second phase (2019-20) can be attributed to physical phenomena such as upwelling processes. The term "upwelling" pertains to the vertical transport of nutrient-laden waters from the deep regions of the ocean to its surface, leading to heightened levels of primary productivity and chlorophyll-a concentration. The process of upwelling facilitates the upward movement of nutrient-laden waters from lower depths to the uppermost layers, thereby serving as a vital nutrient supply for the proliferation of phytoplankton. The increase in nutrient levels serves as a catalyst for the proliferation of phytoplankton, resulting in elevated concentrations of chlorophyll-a. Coastal regions are notably impacted by upwelling phenomena as a result of the presence of coastal upwelling systems. These systems facilitate the upwelling of nutrient-rich waters in coastal regions, resulting in elevated levels of chlorophyll-a and enhanced primary productivity. Numerous investigations [96-98] have been conducted to explore the relationship between the intensity of upwelling and the concentrations of chlorophyll-a in various geographical areas. These studies have yielded compelling evidence supporting the impact of upwelling on primary productivity and the dynamics of chlorophyll-a.
To assess the overall water quality, the Water Quality Index (WQI) was calculated for different zones in different phases (2013-15, 2015-17 and 2017-20) as well as for the whole period. Figure 4 and 5 show the values of different zones in relation to the WQI. The total value of coastal waters fell below Class C in 2017-20, indicating the impact of river inflows, port activities and industrial discharges in and around Paradeep. This finding clearly indicates that the water quality of the Mahanadi River and its estuary falls below Class D. The WQI ratings for the estuary zone and mixed zone were 30 and 46, respectively, indicating the impact of loading in the estuary, which may have been diluted beyond the confluence.
In 2013-15, the WQI scores for the Estuarine Zone (E), Mixing Zone (MZ), Mixing Zone North (MZN), and Mixing Zone South (MZS) were 44, 47, 65, and 59, respectively. In the subsequent 2017-20 period, these scores dropped to 32, 39, 45, and 38. This decrease, particularly in the estuary zone (from 44 in 2013-15 to 30 in 2017-20), indicates a 32% deterioration in water quality. Although the mixed zone status declined from 47 in 2013-15 to 39 in 2015-17, it improved again to 46 in 2017-20. The mixed zone also showed a downward trend, with the WQI 65 in 2017 drop to 43 in 2020, a 32% decrease. The mixed zone south, influenced by the Jatadhari River and IOCL activities, saw a decline from a WQI score of 59 to 38 (36%) in 2015–17 and further to 41 (30%) in 2017-20. Rainwater, which carries sediment, nutrients and heavy metals, is a major contributor to the deterioration of the Mahanadi River estuary. In addition, the presence of port stockpiles with ore dumps, extensive storage sites for industrial waste (e.g. IFFCO and PPL gypsum), discharges from boats at the fishing dock, agricultural effluents and industrial discharges from Paradeep and the catchment area aggravate the decline in water quality in the estuary and in the mixed zone (class C or D). Therefore, appropriate mitigation plans are required to regulate the flow of pressures and improve ecosystem quality.
A comprehensive analysis was performed considering multiple studies conducted earlier in the same area (Tab.2). In the present study, 18 parameters were examined and various trends analyzed over different study periods. Among the parameters considered was TSS (Total Suspended Solids), turbidity, FC (faecal coliform bacteria), nitrite, phosphate, silicate, mercury and iron were found to be the most important factors for the route studied. These parameters showed significant differences during their respective phases from 2013 to 2020. However, despite their significant contributions, TSS and turbidity remained relatively stable throughout the study period (Fig. 2 a,b,c). During 2017-20 (Fig. 2c), the dominance of iron and mercury in the water samples can be attributed to under management of the Paradeep Harbor ore stock, which allowed stormwater intrusion into the study area, thereby affecting water quality. The increase in nitrite and phosphate levels in the study area over time may be influenced by nearby fertilizer plants such as IFFCO and PPL.
Data from 32 stations monitored from 2013 to 2020 was considered to observe the trends of individual parameters at each station. The regression value (R2) was used to determine the increasing or decreasing trend of each parameter. The percentage of stations showing an increasing or decreasing trend was then calculated and expressed as a percentage of the total 32 stations as shown in Table-5. The aim of this analysis was to provide site-specific trends in pollution levels for important water quality parameters. It underscores the need for constant monitoring to predict pollutant properties, their presence, changes, movement or leakage in complex ecosystems where time is of the essence.
Observed trends indicate that TSS, BOD (Biochemical Oxygen Demand), turbidity and silica are of particular concern, with 44%, 36%, 28% and 28% of stations showing an upward trend respectively. Although high levels of nitrite were expected due to the presence of large fertilizer plants (IFCCO and PPL) nearby, 40% of the stations experienced a decrease. This decrease is due to strict regulations by the regulator on industrial discharges and modernization of household and ETP (Waste Water Treatment Plants) plants in industry.
To characterize the variability of the parameters, a multidimensional scaling analysis (MDS) was performed with IBM SPSS (version: 22). The data set from 2013 to 2020 was subjected to this analysis and revealed four different groups of parameters with similarities (Fig.6). TSS and pH formed a separate group while nitrite, temperature and mercury merged into another group. Another group formed fecal coliforms, turbidity, silicates, phosphates and DO (dissolved oxygen). Chl-a (chlorophyll-a), BOD, TOC (Total Organic Carbon), salinity, iron (Fe) and cadmium (Cd) formed the third distinct group. Remarkably, metals such as manganese (Mn) and lead (Pb) did not belong to any group with a strong similarity. However, in the cluster analysis (Fig.7) three distinct groups were observed with a similarity of 48.2%. Group one included temperature, salinity, Cd, Hg, BOD, TOC and Chl-a. Group two consisted of Mn, Fe and Pb. Group three included parameters such as pH, phosphate, turbidity, nitrate, TSS, silicate, FC and DO, which were partially consistent with MDS analysis. Mn and Pb, which were independent of each other in MDS, showed strong resemblance to Fe in cluster analysis and formed a group. At ~70% similarity level, a narrower microgroup consisting of pH, PO4, turbidity and nitrate emerged, which showed differences to TSS, silicate, FC and DO within the third group. This clearly demonstrates the interdependence of parameters such as DO, PO4, silicate, TSS, turbidity and nitrite, which control ecosystem biological community parameters and consequently affect coastal health. Likewise, BOD and chlorophyll-a are interdependent and play a crucial role in the health of the coastal environment [31, 47, 99]. Hierarchical cluster analysis (Fig.7) shows three distinct groups, with metals (Fe, Mn and Pb) forming a separate group [100]. Cd and Hg, which share similarities with salinity, are in a higher range of distinguishable similarity [47]. BOD, TOC and Chl-a are grouped together but differ significantly at a higher similarity range and are closely related [101]. Phosphate, DO, silicate, TOC, turbidity, FC and nitrite are also grouped together, indicating a higher degree of similarity compared to other groups [102].
In the multi-dimensional scaling (MDS) analysis (Fig.8), six distinct groups were identified upon comparing various stations. The aforementioned groups are linked to four distinct zones, which are referred to as the Estuary (E), Mixing Zone (MZ), Mixing Zone South (MZS), and Mixing Zone North (MZN). It is worth mentioning that the estuary stations, ES1 to ES6, constituted a distinct group. Further, MZ9 of M zone and MZS7 of MS zone exhibited a close association with the E zone. The mixing zone stations displayed a range of distinct characteristics and were classified into various categories. Within the set of Mixing Zone stations, MZ2 is associated with the MZN zone, MZ1 is affiliated with the MZS Zone, and MZ5 has formed a distinct new group with MZS4 of MZS zone. Upon analysis of the MN zones' stations, it was observed that MZN3, MZN4, MZN5, and MZN6, in addition to MZ2 of the Mixing Zone, exhibited a discernible cluster. Nevertheless, MZN1 and MZN2 exhibited a noticeable clustering pattern with MZ8 of the Mixing Zone. The stations within the Mixing Zone South, have been classified into four distinct categories. Specifically, MZS11, MZS3, and MZS9 have been grouped together, while MZS1, MZS6, MZS10, MZS2, MZS5, and MZS9 have been grouped with the mixing zone. Eventually, MZS4 of the Mixing Zone South zone and MZ5 of the Mixing Zone separated into their own distinct group.
The principal component analyses were performed on the parameters of coastal water samples from 2013 to 2020. The results are presented in Table-6, which identifies the correlated variables (principal components) from the extensive data set. The table includes the loadings and eigenvalues for each component, as well as the percentage and cumulative percentage of variance explained by each parameter. The objective is to explain the maximum variance with the fewest number of principal components.
The analysis revealed that the first eight principal components account for over 90% of the total variance in the dataset. The first principal component accounts for 34.34%, the second for 17.92%, the third for 9.49%, the fourth for 7.83%, and the fifth for 7.02% of the total variance. The eigenvalues of the first five principal components can be used to assess the dominant physicochemical and biological processes of the coastal ecosystem.
In the first principal component (PC1), salinity, turbidity, BOD, TOC, phosphate, silicate, Chl-a, Mn, Fe, and Cd showed strong positive loadings (0.516-0.823), indicating an interdependence between phytoplankton growth, increased chlorophyll levels, and the rise in phosphate and silicate. This, in turn, contributes to the increase in BOD due to the death and decay of phytoplankton [103]. This finding supports the results of previous studies [104-106]. Turbidity played a significant role with the highest loading in this study.
The second major component (PC2) exhibited high positive loadings for dissolved oxygen (DO) (0.722) and fecal coliform (FC) (0.691), along with negative loadings for pH (-0.747), total suspended solids (TSS) (-0.549), and salinity (-0.607). The inflow of the Mahanadi River into the study area brings surface water and industrial effluent, which may explain these associations. Salinity showed a negative correlation with DO and silicate, while DO solubility decreased with increasing salinity. Freshwater inflow from rivers is considered the primary source of silicate and phosphate in coastal waters [107]. The positive correlation between DO, phosphate, and silicate in this study could be attributed to the intrusion of freshwater with low salinity but high nutrient content (phosphate and silicate) into the coastal waters [108], thus confirming our findings.
The third (PC3) and fourth (PC4) principal components did not exhibit significant positive or negative loadings for any of the parameters. In PC5, BOD showed a negative loading (-0.649) without any notable evidence of positive loading.
The cumulative percentage of variance for the different components, presented in ascending order (34.38, 52.30, 61.8, 69.63, 76.64, 81.9, 86.4, 90.2, etc.), indicates that the first eight components account for over 90% of the total variance. The first five components explain over 70% of the deviations and are influenced by parameters such as salinity, turbidity, BOD, TOC, phosphate, silicate, Chl-a, Mn, Fe, and Cd, which likely play a crucial role in determining the characteristics of this coastal ecosystem.