Assessment of wastewater treatment potential of sand beds of River Ganga at Varanasi, India

The disposition of the sand bed at the convex side is not only in the case of Varanasi city, but has also been confirmed by other researchers for different river systems (Metcalf & Eddy 1979; Prasad et al. 2006; Alekseevskiy et al. 2008; Trombadore et al. 2020). The above studies show that a huge sand bed is available at the convex side of River Ganga, which can be utilised for wastewater filtration and management, especially in the lean period (discharge is very low and sand bed exposure is maximum). From the literature, it can be shown that the rate of filtration, size of sand particles, depth of filters, and temperature of sand have a considerable effect on wastewater filtration (Al-Adham 1989; Check et al. 1994; Ausland et al. 2002; Hua et al. 2003; Zahid 2003; Prasad et al. 2006; Kumar et al. 2021). In order to choose a better location for the disposal of wastewater in sand beds, it is necessary to find out the efficiency of lateral filtration. Therefore, the work done on lateral flow sand filters (LFSFs), which are close to the field conditions of river sand bed is studied and presented here.

Check et al. (1994) did an experiment on a lateral-flow sand filter (LFSF) system. Three full-size cross-sectional models of the LFSF system, each with a different sand fill, were constructed. The laboratory system represents a slice through the LFSF in the direction of flow (referred to here as the down slope length). The removal efficiencies (%) for different parameters are chloride: 14.0, orthophosphate (P): 39.7, ammonia (N): 99.9, TKN (total Kjeldahl nitrogen): 99.1, nitrogen (N): 22.9, suspended solids: 99.4, TOC: 88.3, BOD: >99.1, TC: 100, and faecal coliform: 100. The treatment differences resulting from the use of three separate sand fills were minimal. The use of a relatively coarser, more permeable sand fill can be recommended, as it allows good treatment as well as enhanced hydraulic functioning and system longevity. LFSF is closer to the structure of the river sand bed, which has a slope towards the river. The effects of the rate of loading and temperature need to be analysed for their effective application in the field. Havard et al. (2008) performed experiments on LFSFs for their treatment of septic tank effluent in Truro, Nova Scotia, Canada. This study also supports the idea of utilising the river sand bed, which has a slope towards the river, and which can be very useful in removing the BOD load.

In the present research study, a 7-km stretch (measured along the bend) of River Ganga at Varanasi has been selected. This stretch has been divided into three zones: entry, middle, and exit zones. The three sections have been selected in the respective zones, i.e. Ravidas Ghat (section 1-1), Dashaswamedh Ghat (section 2-2), and Panchganga Ghat (section 3-3) as shown in Figure 1. The objective of this research is to assess the filtering potential of the selected sections and to find out the most suitable zone among the three zones for wastewater filtration. For fulfilling the objectives, studies were done to find out (a) the zone of the longest filtration length in the transverse direction of the sand bed and the maximum volume of sand available for wastewater filtration and (b) the difference in the filtration efficiency of the three different sand columns (filled with sand from the selected three zones) on the basis of their ability to vary the following parameters: DO, BOD, EC, TDS, temperature, and pH of the wastewater.

To fulfill the first objective, a morphological survey and bathymetrical survey of the three zones were carried out and the methodology is presented in the first section. For the second objective, measurement of the physical characteristics of the sand samples and experimental work on the filtration of secondary-treated effluent through the three different vertical sand columns were done and presented in the second and third sections, respectively.

Selected sections of River Ganga and sand bed were thoroughly surveyed. The river section's depth, width, and RL (reduced level) of river water were determined. The width and RL of the sand bed, the longitudinal distance between the sections, and the distance of the sand bed's maximum elevation from the concave bank are determined and presented in Table 2. The data were collected for the month of April 2017 (a period of low discharge). The geographical coordinates of the sections were obtained from Google Earth's satellite image (presented in Table 1) and the locations of the sections are presented in Figure 1.

Measurement of the RL of the water level of the selected sections: the RL is marked on a permanent stable structure above the water level at Shiwala Ghat and at Vijay Nagar Ghat. These ghats are situated on the concave bank of River Ganga at Varanasi. At both ghats, the marked RL was transferred onto the water level with the help of theodolite and the staff. The distance between these two ghats was measured and the slope of the water surface was calculated. The RL of the concave side of the three sections (sections 1-1, 2-2, and 3-3) was calculated by calculating the distance of the section from Shiwala Ghat.

To observe and compare the treatment performance of the collected sand samples from the three sections, three cylindrical filters of diameter 15 cm and height 45 cm were constructed. These filters have perforated bottoms to pass out the effluent. Sand samples were filled up to 30 cm height in each cylinder. The three cylindrical containers had conical funnels and containers to collect the filtered effluent. Figure 3 shows the schematic sketch of the filter.

Secondary-treated effluent was brought from the Bhagwanpur sewage treatment plant (STP) daily and dosed 2 l to each filter at the rate of 0.15 m³/m²/h intermittently once in a day, for 45 days, to mature the filters. The time for the filters to stabilise and reach a constant performance (maturation) was found to be approximately 10 days, but progression to maturity up to 40 days was also observed (Bauer et al. 2011). From the experience of earlier published works, an initial 45-day period was utilised for the maturity of sand filters in the present work. The maturity of sand bed means sufficient growth of microbes, which is responsible for the organic matter removal. The same process continued and sampling of the treated effluent was started after 45 days of maturity. It took 45 min for the filtration of wastewater, after which, sand columns were kept as such in the open atmosphere for the rest of the day. Samples were collected for 50 days at a 5-day interval. DO, BOD, EC, TDS, salinity, pH, and temperature of the effluent were measured before and after the filtration through sand columns. DO, EC, TDS, salinity, and pH were measured with the digital probe of HACH, USA. BOD was measured by the standard method (APHA 1998) and temperature was measured by a glass thermometer.

The results are presented in three parts: (1) Results of morphological and bathymetrical surveys of the selected sections, which are (a) locations of sand beds; (b) width, maximum height, average slope, and cross-sectional area of sand beds of the three selected sections that give the zone of the longest filtration length and the maximum volume of sand available for wastewater filtration. (2) Results of the measurement of physical characteristics of the sand samples of the three sections, which are D10, Cu, and K that are used in further analysis of the effect of size and composition of sand on wastewater filtration. (3) Results of the measurement of filtration efficiency of three different sand columns, which give differences in the filtration potential of the sand columns.

Survey with satellite imageries: The satellite imagery of the sand bed (Figure 1) of River Ganga clearly shows that the width of the sand bed is the maximum in the middle region of the bend. The data in Table 1 are taken from Google satellite imagery for the month of April 2017. The RL of the concave bank first increases from section 1-1 to section 2-2, and then, it decreases as one goes further downstream from section 2-2 to section 3-3. Figure 1 and Table 1 indicate that the river is curvilinear both in horizontal and vertical planes. The RL of the convex bank, i.e. sand bed increases continuously from section 1-1 to section 3-3. Table 2 shows the morphological characteristics (average slope, height, width, cross-sectional area) of the sand bed at the selected sections. The data indicate that the height of the sand bed increases as we move downstream in the bend.

As per Indian Standard (IS) code 383:1970, natural sand can be defined as a fine aggregate (particle size is less than 4.75 mm) produced by the natural disintegration of rock and which has been deposited by streams or glacial agencies. Sieve analysis results suggested that the sand deposited in all three sections is fine sand with an effective diameter (D10) of 0.137, 0.097, and 0.129 for sections 1-1, 2-2, and 3-3, respectively. The characteristics of fine sand in these sections are almost similar. Fine sand is hard and durable in nature, very clean, and free from adhering coatings and organic matter. The coefficient of permeability (K) for all three classes has been calculated from a constant head permeability test. The value of the coefficient of permeability (K) for section 1-1 is the highest (3.75 × 10⁻² cm/s), while for section 2-2, it is the lowest (1.27 × 10⁻² cm/s) as shown in Table 3.

In order to find out the filtration efficiency of sand in these three sections, the effluent is taken out through three cylindrical filters of the same size (diameter as well as height) containing sand samples from sections 1-1, 2-2, and 3-3. About 2 l of secondary-treated effluent was dosed intermittently through filters once in a day for 45 days and the filtered sample was collected and analysed at a 5-day interval. An intermittent loading rate of 2.67 l/h (0.15 m³/m²/h) is maintained for all three sand samples throughout the experiment. Table 4 shows the mean, median, maximum, minimum, and standard deviation values of the parameters before and after filtration. Treatment performance (% increase and removal) of filters is calculated using mean values of the parameters.

Cross-sectional details of the selected sections are presented in plots between depths of river/elevation of sand bed vs. transverse distance at three selected sections of River Ganga at Varanasi in Figure 2(a)–2(c) and in Table 2. These data are obtained by field observations through tacheometry.

The width of the river channel at Ravidas Ghat is 398 m. The formation of the sand bed has been well pronounced at the convex side. The width of the sand bed is 563 m. Slope and cross-sectional area are minimum at this section, among the selected three sections.

From the tacheometric survey and data of satellite imagery, we got similar results, which means, as we move downstream in a bend, the height of the sand bed increases. Among the three selected sections, the width and cross-sectional area of the sand bed are maximum at the middle region, i.e. at section 2-2 (Dashaswamedh Ghat). Therefore, at section 2-2, the maximum amount of sand is available for wastewater filtration. And the maximum filtration length is also available in this section.

At the entry region of the bend, the curvature of the bend is less; therefore, the centrifugal force is less, and secondary cells are less pronounced. This may be one of the reasons why the width of the sand bed is less here. As the flow moves downstream towards the middle region of the bend, centrifugal forces enhance and lead to well-developed secondary cells (Choudhary 1974). Enhancement in centrifugal forces may be one of the main causes for the increase in the width of the sand bed in the middle region of the bend. In the exit region again, centrifugal forces reduce and the width of the sand bed reduces. The height of sedimentation (sand bed) is maximum at section 3-3 and it reduces as we move upstream towards section 1-1. This is because of the formation of a separation zone in zone 3, which leads to the maximum height of sedimentation in this zone.

From Table 3, it is clear that although all three sand samples are in the fine sand category, their D10 values vary from section to section. At the entry zone of the bend, coarser sand is deposited, as the flow moves further towards the central zone (section 2-2), the sand becomes finer. Again, the result of sieve analysis shows that the sand becomes coarser in the exit region as compared to that of the section in the middle region. However, it is observed by the naked eye that there is a mixture of sand and fine silt at section 3-3. So, maybe, the effective diameter D10 of the sand of section 3-3 by sieve analysis gives a little higher value than its actual value. This may be due to the loss of fine silt during sieve analysis.

As the flow traverses from an upstream bend to a downstream bend, the concave bank converts into the convex bank and vice-versa. It means that the bank where erosion occurs in the upstream bend is slowly converted to the bank where sedimentation happens in the downstream bend (Omar et al. 2022b). The orientation of secondary cells reverses. The residual effect of the upstream bend does not allow smaller sand particles to settle down in the entry region of the downstream bend. As the flow traverses towards the middle region of the downstream bend, centrifugal forces enhance and the residual effect of the upstream bend reduces. In this region, finer sand particles also deposit at the upper layer of the sand bed. In the exit region of the bend, streamlines diverge and a separation zone is formed at the convex side. Hence, silt and clay are also deposited with sand in this region.

To measure the treatment performance of the sand samples, the data of filtration through vertical sand filters are discussed here. Treatment performances of the three sand samples are compared under the same atmospheric conditions and loading rates. The height of the sand column is kept constant for all three sand samples throughout the experiment.

The percentage increase in the mean value of DO is the highest for the sand sample of section 3-3 and the lowest for section 1-1. It shows that the sand sample of section 3-3 is better than the other two in terms of DO increase.

Percentage removal in the mean value of BOD is the highest for the sand sample of section 3-3 and the lowest for section 1-1. The difference in the percentage removal for sections 2-2 and 3-3 is very small; therefore, BOD removal from the sand samples of sections 2-2 and 3-3 is nearly equal. The higher BOD removal from the sand of the central zone (zone 2) and the exit zone (zone 3) than the entry zone (zone 1) is attributed to finer sand deposition in the central zone, and silt along with sand in the exit zone. Table 4 also shows that the average removal of BOD from the sand of section 2-2 was 93.7% when the hydraulic loading rate was maintained at 0.15 m/h. Al-Adham did an experiment at a nearly equal hydraulic loading rate of 0.16 m/h and found the average removal of BOD to be 86% for a 20-day filtration run. The removal percentage is higher than Al-Adham's results, which may be because he performed the experiments in the winter season (at a lower temperature).

The value of percentage removal in the mean values of EC and TDS is 4.9, 3.4, and 2.3% for sections 1-1, 2-2, and 3-3, respectively. It shows that there is a very slight variation in the EC and TDS of secondary-treated effluent after sand filtration. This removal occurs due to the adsorption of sand particles.

The mean value of daywise variation in the pH of the influent and the effluent shows that it changes from slightly acidic to slightly alkaline after filtration for all sand filters.

Data of each day show that the temperature of the effluent after filtration is more than the temperature of the influent for all three filters, and the percentage increase in the mean value of temperature is nearly the same for all three filters.

Selection of the best section for the filtration of wastewater: Among the three sections, the width and approximate cross-sectional area of the sand bed at section 2-2 is maximum and their values are 856 m and 1,404.4 m², respectively, for the month of April 2017 (lean period). The maximum filtration length and the maximum volume of sand for the filtration of wastewater are available in zone 2.

Percentage BOD removal by vertical sand filters made by sands of sections 1-1, 2-2, and 3-3 are 91.2, 93.7, and 94.0, respectively. In terms of BOD removal, the sand of section 3-3 is slightly better than the other two sand samples. The reason may be the presence of silt along with the sand in zone 3 (section 3-3).

The percentage of BOD removal is slightly more for the sand of section 3-3; however, due to the formation of a separation zone in the stream in this region, pollutants are likely to be stagnant and not easily dispersed in this region. Hence, section 3-3 is not suitable for wastewater disposal. Therefore, in view of the above results, section 2-2 seems to be the better section for wastewater disposal.

Sand samples of River Ganga have been collected from three different zones and an attempt has been made to represent an entire zone (1, 2, and 3) from that sample. An experimental study can be carried out to select more suitable sites to obtain more useful results in real-field conditions. More laboratory tests can be conducted to simulate the actual morphology of the sand bed and to evaluate the effect of other hydraulic parameters that affect the flow of wastewater into the natural river sand bed.

The height of the sand bed continuously increases as the distance downstream increases and it attains the maximum at the end of the bend. Among the three selected zones in the river bend, the filtration length and the volume of the sand available are maximum in the central zone (zone 2) and their values are approximately 856 m and 14.04 × 105 m³, respectively. The effective diameter (D10) and the coefficient of permeability (K) of the sand at the central zone are the lowest among the three zones. Hence, out of the three selected zones, the middle zone of the sand bed has the best sand deposits. The sand sample of the exit zone (zone 3) shows that a small amount of silt is also deposited with the sand in this zone. The sand filters 1, 2, and 3 are quite efficient in organic load (BOD) removal with mean efficiencies of 91.2, 93.7, and 94.0%, respectively. The performance of sands of zones 2 and 3 is nearly equal and better than the sand of zone 1. The sand filters have lesser mean removal efficiencies of 2–5% for EC and TDS. The pH of wastewater slightly changes from acidic to alkaline with the values for sand filters 1, 2, and 3: it changes from 6.8 to 7.38, from 6.8 to 7.39, and from 6.8 to 7.47, respectively. The mean temperature of 26.18 °C of the wastewater increases to 30.3, 30.5, and 29.6 °C after the filtration from sand filters 1, 2, and 3, respectively. This suggests that the sand temperature is higher than that of the wastewater but this higher temperature is almost the same for all three filters. Although BOD removal is the function of the rise in the temperature of the sand bed, the rise in temperature is nearly the same for all three filters. Therefore, the observed differences in the sand filters in BOD removal are mainly due to the difference in the particle size of the sand. From the interpretation of the above conclusions, it is clear that the convex side of the sand bed of the central zone (zone 2) of the river bend is found to be the most suitable for the disposal and filtration of wastewater.

All relevant data are included in the paper.

The authors declare there is no conflict.