Composite radial filter for removal of agrochemicals from agricultural runoff Open Access

Groundwater is the major source of requirements for ever-growing domestic, agricultural and industrial sectors resulting in lowering of groundwater levels. The water table in Punjab is declining at an alarming rate of 55 cm annually in the year 2017 (Anon 2020). It is necessary to recharge groundwater aquifer with available surface water for managing groundwater resource. The excess canal water and runoff water flowing from agricultural fields could be utilized for the augmentation of groundwater. However, agricultural runoff water may contain agrochemicals particularly N, P and K which contaminate groundwater aquifer. Agriculture is major source of non-point source (NPS) contamination of surface as well as ground water. The NPS pollutant available in agricultural runoff should be removed before recharging groundwater aquifer. Farmlands are one of the major non-point nutrient sources because of the increased use of chemical fertilizers in present day. Nitrogen from paddy fields runoff needs to be examined for recharging aquifers. Runoff nitrogen from paddy fields was evaluated by the various researchers (Kuroda et al. 1995; Jeon et al. 2004). Yoshinaga et al. (2007) examined nitrogen load from a large-sized paddy field during the crop period and the results of field measurements of nitrogen load were found to be 18.8 kg ha⁻¹, with 7.2 kg ha⁻¹ from surface drainage and 11.6 kg ha⁻¹ from percolation loss. The mechanisms of agricultural non-point pollution processes are very complex and various models are used to estimate non-point source pollution by nitrogen and phosphorus (Xu & Zhang 2006). Phosphorus and nitrogen in agricultural land can be taken up by plants, bound by clay particles, can infiltrate into soil profile and be drained out as surface runoff and vitalize into atmosphere. The annual average values of ammonium nitrogen in farmland drainage of Yixing and Wangzhuang China were found to be 1.596 and 2.103 mg L⁻¹, respectively (Chen et al. 2010). Sharma et al. (2019) reported nitrogen (N), phosphorus (P) and potassium (K) content in the runoff water from rice and wheat fields were 33.7, 5.5 and 20.0 ppm and 33.3, 6.5 and 20.0 ppm respectively, which are above the permissible limit. The physico-chemical quality parameters of Ganaga canal water near Haridwar (Uttrakhand) and Allahabad (UP) regions were found within the permissible limit for irrigation (Verma et al. 2012; Matta et al. 2015). The physico-chemical and biological parameters, specifically, biological oxygen demand (BOD), chemical oxygen demand (COD), TDS and most probable number (MPN), in Phuleli canal water at Hyderabad were found to be more than permissible limits (Channa et al. 2015). Water in a rice field is stagnant and may contain many pathogens and other biological contaminants. Monitoring of the biological properties of water relies largely on examination of presence and absence of indicator bacteria (coliforms and Escherichia coli). The pathogen communities that co-occur or occur in absence of indicator bacteria need to be assessed. Total coliforms, E. coli and emerging pathogen species of Aeromonas, Yersinia, Salmonella, Shigella, Klebsiella, Staphylococcus, Listeria, Vibrio and Campylobacter could be detected by Bacteriological Water Testing Kit (BWTK) (Sahota et al. 2010). Kaur & Mehra (2012) reported coliform contamination in the Yamuna river water flowing through Delhi city at Wazirbad barrage, Nigambodh ghat and Okhla barrage during the year 2010 and reported that Okhla barrage has maximum of total E. coli bacteria followed by Nigambodh ghat and Wazirabad barrage due to industrial as well as anthropogenic activities.

The artificial recharge technique with efficient filtration unit is the best way to augment the groundwater. Filtration unit has to be designed based on pollutant available in surface water and efficient filter media must be provided which could remove the physico-chemical particularly N, P, K from agricultural runoff water. The frequent clogging of filtration media due to suspended solids in the surface water reduces the efficiency and effectiveness of the filter. Hydraulic efficiency and effectiveness are the important measures for any filter. An optimum design of a filtration unit is based upon various factors such as quantity and quality of recharging water, soil-climate-aquifer characteristics, agro-chemicals, vegetation, population and location (Gaurav et al. 2019). The filtration system can be classified as vertical, horizontal and radial based on flow direction. There are various advantages of radial filter over vertical and horizontal flow filters due to increase in flow path increases, which cause more resistance, offering more surface area, opportunity time and pollutant adsorption on filter materials. The siltation settles down on the floor of the filtration unit instead of filtering media, which results in less choking and clogging problems (Gaurav et al. 2019).

There are many filter materials such as brick flakes, coarse sand (CS), gravel (G), zeolite (Z), charcoal, pebbles (P), alum, etc., which can be used for filtration of water. An experiment was conducted on filtration unit consisting of boulders, stones, charcoal and sand for recharging through well at Bengalore, Karnataka and found the quality of recharged water within the permissible limits (Ramachandrappa et al. 2015). Kambale et al. (2009) determined the optimum thickness ratio of filter materials viz. CS, G and P used in a filtration unit installed in a recharge shaft. Five different thickness combinations of CS, G and P were evaluated in laboratory using artificially synthesized water based on concentration level of different impurities present in natural storm water and concluded that thickness ratio of 1.53 (CS:P) found to be the best in all five combinations. Performance of rectangular filter of size 200×160×1,200 mm consisting of CS and P was evaluated for recharge rate, filtration efficiency and clogging time by varying thickness of CS and sediment concentration (Kumar & Singh 2014). The hydraulic performance and physical characteristics of filter media viz. Z, scoria, sand and polymeric beads were evaluated for flow-through rates and clogging times through column study and observed that scoria-based filter was found to be highly variable in performance and Z with low flow rate had highest treatment efficiency of 88% among all the filter media (Kandra et al. 2014). Zeolites have large porosity, cation exchange capacity and selectivity for ammonium and potassium cations (Sangeetha & Baskar 2016). The reuse of filtered water from the filtration unit using bioassays was evaluated and removal efficiency of toxicity found to be from 56% to 99% through ozonation (Zhanga et al. 2011). A filtration device including constructed wetland, settling pond and porous recharge reservoir, which receives surface runoff water through constructed vegetative wetland, goes to a porous reservoir that traps sediments and pollutants. A back flushing arrangement was made in the porous reservoir to clean trapped sediments and particulates (Czarnecki 2012). Jaskunas et al. (2015) conducted experiments on adsorption of potassium on clinoptilolite after 7 hours fast start, becoming diffusion controlled and slowing down. The rate and capacity of adsorption do not depend on type of solution used but favor higher temperatures. The applicability of Languir and Dubininn-Radushkevich models indicated a non-uniform nature of the clinoptilolite surface and suggested formation of an adsorbate mono layer, which was used to calculate active surface area clinoptilolite. Gaurav et al. (2019) developed a composite radial filter consisting of activated charcoal (C), CS and G filter medias in thickness combination ratio of 1:2:2 (10, 20 and 20 cm) for the removal of impurities from canal water for recharging aquifer and the water quality parameters of filtered water were found to be within the permissible limit at lower flow rate (0.3 lps).

The excess runoff from rice fields is available particularly during heavy rainfall that may be utilized for recharging a groundwater aquifer after proper filtration of bacteriological contamination and agrochemicals, particularly N, P and K, in runoff water. The present study was undertaken with the objective to develop and evaluate a filter for the removal of agrochemicals and bacterial contamination from agricultural runoff.

The composite radial filter was made of 8 gauge GI sheet in cylindrical shape of height 45 cm and circular base of radius 100 cm (Gaurav et al. 2019). Concentric rings of iron mesh with various radius of height 45 cm wrapped with nylon net were fabricated and fixed on the base of GI sheet at different interval for various combination treatments (Figure 1) The filter media viz. brick flakes (BF), CS, C and Z were selected (Figure 2) based on the literature (Dalahmeh 2013; Jaskunas et al. 2015; Gaurav et al. 2019) and placed in annular concentric ring in sequence of outer to inner ring for the construction of composite radial filter. The hydraulic conductivity of G, CS, C and Z materials was determined experimentally using constant head permeameter. The experiment was carried out with synthetic N, P and K solution through column study to get the optimum thickness of the filter media for the removal of N P K. The size of the filter media viz. G (2–4 mm), CS (1.18–2 mm) and C (2–4 mm) were used (Kambale et al. 2009; Gaurav et al. 2019) at different thickness combination ratios, viz. of (Z:C:CS:G)=1:2:4:4 (T₁), 1:3:4:4 (T₂), 1:2:2:4 (T₃), 1:3:2:4 (T₄) and 0:2:1:2 (T₅) respectively, where T_1, T_2, T_3, T_4, T₅ are the various treatments (Table 1). The height of the filter materials (C, CS and G) were placed up to a height of 30 cm each (design depth of filter media) while Z was kept up to 16 cm height (actual design depth) due to the cost of the material in consideration. The BF were placed in outermost concentric ring to check trash, litter, foreign materials, sediments, etc. in runoff water from entering into the composite radial filter.

The chlorination (Figure 1) unit consists of a plastic box with a rod placed in the inner area of the innermost concentric ring from the centre of filtration unit and opening the cannula discharges solution into the water. As per the study, about 0.1 ml of 5% sodium hypochlorite solution is required for 1 L of water (Singh 2015). The supply rate of sodium hypochlorite solution through the cannula was calibrated with different discharge rates at the outlet of the filter. The sodium hypochlorite solution is supplied with help of a plastic can having a cannula opening, which discharges at the required calibrated rates. The water samples were analyzed simultaneously for physico-chemical and biological parameters such as N, P, K, BOD, COD, TDS, pH, MPN etc.

NO₃-N and P were analyzed with a spectrophotometer and K₂O with a flame photometer. The pH and TDS were determined with pH meter and EC meter respectively. The BOD is determined by measuring dissolved oxygen (DO) content of runoff water before and after incubation at 25 °C for 3 days, which was measured by the Winkler method (Winkler 1888). The COD was determined by a titration process in which all the organic compounds could be fully oxidized to carbon dioxide with potassium dichromate (K₂Cr₂O₇), a strong oxidizing agent.

The bacteriological analysis of runoff water samples was carried out for MPN index, total coliforms, faecal coliforms – E. coli by standard methods (IS-10500-2012 BIS, New Delhi, India), which are MPN index method and BWTK for emerging pathogens and diversity of bacteria in rice field water (Pandove & Sahota 2013). The BWTK is an auto analytic, affordable kit and is a modification of the multiple tube method (MPN), providing a premeasured broth in a single dose kit. A single test of a water sample is inoculated in a suitable volume of appropriate liquid media and presence or absence of coliforms and emerging pathogens can be determined by growth and specific changes in medium within 48 hours (Figure 3).

The water sample was aseptically drawn in pre-sterilized glass bottles for MPN analysis of using double strength (DS) and single strength (SS) MacConkey broth tubes (Figure 4). MacConkey purple media of SS and DS was prepared in test tubes with Durham's tube and autoclaved. Three sets of test tubes containing three tubes in each set; one set with 10 ml of DS and the other two containing 10 ml of SS were taken. One ml of water sample from the DS broth tubes was transferred to each of three tubes and incubated at 37 °C for 24 hours. The change in colour of the media was observed in Durham's tube after incubation. The number of positive results from each set was recorded (Figure 5) and compared with the standard chart (WHO 1965) to give presumptive count per 100 ml of water sample. The confirmatory test was performed by streaking the positive tubes on eosin methylene blue agar (EMB), MacConkey agar, UTI Hi Crome agar plates for confirmation of the diversity of pathogens and non-pathogenic bacteria present in the rice field water sample.

A composite radial filter consisting of G, CS and C filter media was designed and developed for removal of agrochemicals particularly N, P, K and bacteriological impurities from rice field runoff water for recharging an aquifer. The hydraulic conductivity of G, CS, C and Z materials was determined experimentally and found to be 2.46, 0.83, 2.20 and 0.89 cm/s respectively.

The removal efficiency for NO₃-N and P through C was found through column study in the range of 73–79.5% and 76–79.5% with thickness of 10–25 cm respectively. However, removal efficiency of K₂O through Z was observed to be 73.75–80% with 5–25 cm thickness. There is little variation in removal efficiency for NO₃-N and P with thickness variation of 10–25 cm and K₂O with thickness of 5–25 cm (Table 2). The thickness of C and Z was considered in the range of 10–20 and 5 cm (fixed) for the removal of NO₃-N, P and K₂O respectively.

The equivalent hydraulic conductivity of the five combination treatments viz. T₁, T₂, T₃, T₄ and T₅ was computed using Equation (1) and found to be 1.96, 2.31, 2.18, 2.39 and 2.59 cm/s respectively (Table 3).

The design depth (h_d) of the composite radial filter for five treatments (T₁, T₂, T₃, T₄ and T₅) was computed using Equation (2) for maximum discharge of 1,660 cm³/s (Table 4) and compared with the equation (Gaurav et al. 2019) h_d=0.007Q+1.8. The maximum design depth (h_d) was computed to be 13.6 cm for T₁ treatment (Table 4) and the value of design depth from relation given by Gaurav et al. 2019 was found to be 13.42 cm which is very closed to computed value. The depth of filter material was taken to be 30 cm considering a freeboard of 15–16 cm to overcome capillary rise and overflow. The height of the circular wire mesh wrapped with nylon net was kept at 45 cm for safety measurement purposes.

The performance of the composite radial filter was evaluated by analysing water quality parameters (physico-chemical including N, P and K and biological parameters) of rice field runoff water before and after filtration at different flow rates for five treatments (T₁, T₂, T₃, T₄ and T₅). The value of NO₃-N, P, K₂O, BOD, COD and MPN of runoff water before filtration was found to be 30.2, 30.6, 30.6, 250, 390 ppm and 1100 MPN respectively, which were much higher than permissible limit of 10, 5, 12, 3, 7 ppm and 10 MPN/100 ml respectively (Table 5). The runoff water was allowed to flow through the developed composite radial filter for the five treatments at four discharge rates viz. Q₁=0.42 lps, Q₂=0.82 lps, Q₃=1.24 lps and Q₄=1.66 lps and filtered water was collected at the outlet of radial filter for analysing its quality parameters. The values of water quality parameters after filtration were found to be within permissible limits for all the discharge rates (Table 5). The minimum value of NO₃-N concentration in filtered water was observed to be 3.1, 3.6, 4.5, 5 ppm (within permissible limit) and P concentration was 3.5, 3.9, 5 and 5.5 ppm (within permissible limit) at discharge rates Q₁, Q₂, Q₃ and Q₄ respectively from 30.2 ppm of NO₃-N and 30.6 ppm of P before filtration for treatment T₅ (Table 5). However, the minimum value of K₂O in filtered water was reduced from 30.6 ppm to 4.8, 4.9, 6.6 and 6.9 ppm at flow rates of Q₁, Q₂, Q₃ and Q₄ respectively for treatment T₂ (Table 5) which is under permissible limit.

The value of BOD and COD in filtered water was also reduced for all the treatments and the minimum value was found to be 16, 19, 22, 29 ppm and 27, 32, 42,45 ppm for treatment T₂ at all flow rates viz. Q₁, Q₂, Q₃ and Q₄ respectively (Table 5). The TDS value of filtered water was increased for all the treatments due to residue of C and found within the permissible limit. The minimum value of TDS was found to be 336, 340, 350 and 355 ppm from 332 ppm for treatment T₁ at all flow rates viz. Q₁, Q₂, Q₃ and Q₄ respectively (Table 5). The bacteriological water quality of rice field runoff water was analyzed through MPN analysis. The chlorination by 5% sodium hypochlorite solution of 5 ppm concentration for up to 18 hours of contact time has made the water bacteriologically safe by improving water quality and reducing MPN value to less than the permissible limit (WHO 2017). The MPN value of filtered water was reduced to <10 from 1100 MPN (Table 5). The pH of filtered water was slightly increased from 7.8 to 8.0–8.3 due to the sodium hypochlorite solution and observed to be within the permissible limit for all five treatments at all four discharge rates.

The value of NO₃-N, P and K₂O concentration of filtered water increases with increase in flow rate for all five treatments (Figure 6(a)–6(c)). The minimum concentration of NO₃-N, P and K₂O was found to be at slower flow rate (Q₁=0.42 lps) due to more retention time for adsorbing pollutant. The minimum concentration of NO₃-N and P of filtered water was found for the T₅ treatment and maximum for the T₁ treatment at all flow rates. However, the minimum K₂O concentration was found in filtered water from treatment T₂ due to more thickness of C with presence of Z and the maximum value of K₂O was obtained from the T₅ treatment because it was without Z, which is within permissible limit (Figure 6(c)).

The value of NO₃-N, P_, K₂O of filtered water from the T₅ treatment was found to be minimum and within permissible limit at lower flow rate Q₁ and increases with increase in flow rate. However, all the values were found within permissible limit up to a discharge rate of 1.66 lps. (Figure 7). Concentration of NO₃-N and P of filtered water was found to be minimum and K₂O was maximum for the T₅ treatment as compared to other four treatments, which is within the permissible limit (Table 5). Therefore, T₅ treatment was found to be optimum among all the five treatments from which the physico-chemical and bacteriological parameters of filtered water was found within the permissible limit. The cost of the composite radial filter for the T₅ treatment without Z filter media was computed to be Rs 47000, which is much less than all other treatments (Table 6). The composite radial filter of the T₅ treatment is recommended for filtration of runoff water from rice fields for recharging aquifer, which is technically viable and economically feasible.

The selected T₅ treatment of composite radial filter was evaluated for removal efficiency of NO₃-N, P and K₂O from rice field runoff water. The concentration of NO₃-N, P and K₂O of filtered water increases with increase in time of filtration (Table 7). The removal efficiency of NO₃-N and P was found to be about 80% after 5 hours of filtration while K₂O removal efficiency was dropped to 65% after half an hour of filtration due to the lower filtration capacity of the C filter material. The filter materials were washed with fresh water after 5 hours of operation and dried under the sun. The washed filter materials were used in the composite radial filter and runoff water was allowed to flow through it. The filtered water sample was collected after half an hour of filtration and the values of NO₃-N, P and K₂O was found to be 5.3, 5.8 and 10.9 ppm respectively, which is slightly more than the value obtained from the initial fresh filter material for the same duration, which is 4.3, 5.3 and 10.5 ppm respectively (Table 7). Therefore, the filter materials after washing could be used for filtration.

Composite radial filter consisting of C, CS and G has been designed, developed and evaluated for the removal of agro-chemicals particularly N, P, K and bacteriological contaminants from rice field runoff water for recharging aquifer. The performance of the developed composite radial filter was evaluated by analyzing N, P, K, TDS, BOD, COD, pH and MPN parameters of rice field runoff water before and after filtration for four discharge rates of 0.42 lps, 0.82 lps, 1.24 lps and 1.66 lps respectively for five possible combination treatments. The T₅ treatment consisting of C, CS and G in a thickness ratio of 22 (20 cm:10 cm:20 cm) was found to be best on the basis of N, P, and K removal efficiency and lowest cost (Rs 47,000/-). The removal efficiency of NO₃-N and P was found to be about 80% after 5 hours of filtration from T₅ treatment while K₂O removal efficiency was low about 65% after half an hour of filtration. The filter materials were washed after 5 hours of filtration and reused in composite radial filter for filtration and the value of NO₃-N, P and K₂O in filtered water was obtained to be 5.3, 5.8 and 10.9 ppm respectively, which were almost same values obtained with the fresh filter materials for the same time of operation. The following conclusions have been drawn from the present study.

• (i)

  The composite radial filter consisting of C, CS and G with thickness of 20, 10 and 20 cm (2:1: 2) respectively was developed, evaluated and found to be the best combination for removal of N, P, and K from rice field runoff water at a flow rate less than 1.66 lps for recharging the aquifer.

• (ii)

  The removal efficiency of NO₃-N and P was found to be about 80% after 5 hours of filtration while K₂O removal efficiency was low at about 65% after half an hour of filtration and decreased bacteriological contamination to <10 MPN per 100 ml from 1100 MPN per 100 ml .

• (iii)

  The filter materials after 5 hours of operation could be washed and reused.

We thank Technology Mission Division, Department of Science & Technology (DST), New Delhi, India for providing financial assistance under the scheme ‘North Indian Centre for Water Technology Research in Agriculture’. We also thank Department of Soil and Water Engineering, College of Agricultural Engineering and Technology, Punjab Agricultural University, Ludhiana, India for providing facility to conduct this study.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.