Inoculation with adapted bacterial communities promotes development of full scale slow sand filters for drinking water production

Gravity-driven filtration through slow sand filters (SSFs) is one of the oldest methods for producing drinking water. As water passes through a sand bed, undesired microorganisms and chemicals are removed by interactions with SSF biofilm and its resident microbes. Despite their importance, the processes through which these microbial communities form are largely unknown, as are the factors affecting these processes. In this study, two SSFs constructed using different sand sources were compared to an established filter and observed throughout their maturation process. One SSF was inoculated through addition of sand scraped from established filters, while the other was not inoculated. The operational and developing microbial communities of SSFs, as well as their influents and effluents, were studied by sequencing of 16S ribosomal rRNA genes. A functional microbial community resembling that of the established SSF was achieved in the inoculated SSF, but not in the non-inoculated SSF. Notably, the non-inoculated SSF had significantly (p < 0.01) higher abundances of classes Armatimonadia, Elusimicrobia, Fimbriimonadia, OM190 (phylum Planctomycetota), Parcubacteria, Vampirivibrionia and Verrucomicrobiae. Conversely, it had lower abundances of classes Anaerolineae, Bacilli, bacteriap25 (phylum Myxococcota), Blastocatellia, Entotheonellia, Gemmatimonadetes, lineage 11b (phylum Elusimicrobiota), Nitrospiria, Phycisphaerae, subgroup 22 (phylum Acidobacteriota) and subgroup 11 (phylum Acidobacteriota). Poor performance of neutral models showed that the assembly and dispersal of SSF microbial communities was mainly driven by selection. The temporal turnover of microbial species, as estimated through the scaling exponent of the species-time relationship, was twice as high in the non-inoculated filter (0.946 ± 0.164) compared to the inoculated filter (0.422 ± 0.0431). This study shows that the addition of an inoculum changed the assembly processes within SSFs. Specifically, the rate at which new microorganisms were observed in the biofilm was reduced. The reduced temporal turnover may be driven by inoculating taxa inhibiting growth, potentially via secondary metabolite production. This in turn would allow the inoculation community to persist and contribute to SSF function.

Development of nature-based sustainable passive technologies for treating and disinfecting municipal wastewater: Experiences from constructed wetlands and slow sand filter

There is an urgent need to develop low-cost technology for effective wastewater treatment and its further disinfection to the level that makes it economically useful. This work has designed and evaluated the various types of constructed wetlands (CWs) followed by a slow sand filter (SSF) for wastewater treatment and disinfection. The studied CWs were, CWs with gravels (CW-G), free water surface-CW (FWS-CWs), and CWs integrated microbial fuel cell (MFC) with granular graphite (CW-MFC-GG) planted with Canna indica plant species. These CWs were operated as secondary wastewater treatment technologies followed by SSF for disinfection purposes. The highest total coliform removal was observed in the combination of CW-MFC-GG-SSF which achieved a final concentration of 172 CFU/100 mL, whereas faecal coliform removal was 100 % with the combinations of CW-G-SSF and CW-MFC-GG-SSF, achieving 0 CFU/100 mL in the effluent. In contrast, FWS-SSF achieved the lowest total and faecal coliform removal attaining a final concentration of 542 CFU/100 mL and 240 CFU/100 mL, respectively. Furthermore, E. coli were detected as negative/absent in CW-G-SSF and CW-MFC-GG-SSF, while it was positive for FWS-SSF. In addition, the highest turbidity removal was achieved in CW-MFC-GG and SSF combination of 92.75 % from the municipal wastewater influent turbidity of 82.8 NTU. Furthermore, in terms of overall treatment performance of CW-G-SSF and CW-MFC-GG-SSF, these systems were able to treat 72.7 ± 5.5 % and 67.0 ± 2.4 % of COD and 92.3 % and 87.6 % of phosphate, respectively. Additionally, CW-MFC-GG also exhibited a power density of 85.71 mA/m3 and a current density of 25.71 mW/m3 with 700 Ω of internal resistance. Thus, CW-G and CW-MFC-GG followed by SSF could be a promising solution for enhanced disinfection and wastewater treatment.

Efficiency of a multi-barrier household system for surface water treatment combining a household slow sand filter to a Mesita Azul® ultraviolet disinfection device

Low-cost household technologies for water treatment are crucial to improving drinking water quality and preventing health, social and economic impacts, mostly in middle- and low-income regions. This work assessed the removal efficiency of physical-chemical and bacteriological parameters from river water by a multi-barrier household water treatment system for 113 consecutive days. This system combines a pre-treatment step through a non-woven synthetic blanket, filtration by an intermittent household slow sand filter (HSSF) and a Mesita Azul® ultraviolet disinfection device. In general, the water quality was improved by the evaluated system. Turbidity was removed by an average of 73% (ranging from 33 to 94%), total coliforms (TC) of 3.88 log₁₀ (ranging from 2.22 to 5.16 log₁₀) and E. coli of 2.49 log₁₀ (ranging from 1.81 to 3.30 log₁₀). Filtration improvement was mostly correlated to HSSF biofilm development and influent water quality. Characterisation of HSSF schmutzdecke demonstrated a predominance of organic content, and a higher presence of carbohydrates than proteins on the sand and the blanket. Ultraviolet disinfection with Mesita Azul® inactivated most of the remaining bacteria after filtration and no regrowth was observed after 15 days of disinfection. In conclusion, the multi-barrier household water treatment system was efficient in treating river water, reducing risks of microbial contamination to achieve safe drinking water.

Bioaugmentation of pilot-scale slow sand filters can achieve compliant levels for the micropollutant metaldehyde in a real water matrix

Metaldehyde is a polar, mobile, low molecular weight pesticide that is challenging to remove from drinking water with current adsorption-based micropollutant treatment technologies. Alternative strategies to remove this and compounds with similar properties are necessary to ensure an adequate supply of safe and regulation-compliant drinking water. Biological removal of metaldehyde below the 0.1 µg•L^-1 regulatory concentration was attained in pilot-scale slow sand filters (SSFs) subject to bioaugmentation with metaldehyde-degrading bacteria. To achieve this, a library of degraders was first screened in bench-scale assays for removal at micropollutant concentrations in progressively more challenging conditions, including a mixed microbial community with multiple carbon sources. The best performing strains, A. calcoaceticus E1 and Sphingobium CMET-H, showed removal rates of 0.0012 µg•h^-1•10⁷ cells^-1 and 0.019 µg•h^-1•10⁷ cells^-1 at this scale. These candidates were then used as inocula for bioaugmentation of pilot-scale SSFs. Here, removal of metaldehyde by A. calcoaceticus E1, was insufficient to achieve compliant water regardless testing increasing cell concentrations. Quantification of metaldehyde-degrading genes indicated that aggregation and inadequate distribution of the inoculum in the filters were the likely causes of this outcome. Conversely, bioaugmentation with Sphingobium CMET-H enabled sufficient metaldehyde removal to achieve compliance, with undetectable levels in treated water for at least 14 d (volumetric removal: 0.57 µg•L^-1•h^-1). Bioaugmentation did not affect the background SSF microbial community, and filter function was maintained throughout the trial. Here it has been shown for the first time that bioaugmentation is an efficient strategy to remove the adsorption-resistant pesticide metaldehyde from a real water matrix in upscaled systems. Swift contaminant removal after inoculum addition and persistent activity are two remarkable attributes of this approach that would allow it to effectively manage peaks in metaldehyde concentrations (due to precipitation or increased application) in incoming raw water by matching them with high enough degrading populations. This study provides an example of how stepwise screening of a diverse collection of degraders can lead to successful bioaugmentation and can be used as a template for other problematic adsorption-resistant compounds in drinking water purification.

Performance of geotextile-based slow sand filter media in removing total coli for drinking water treatment using system dynamics modelling

In a slow sand filter, a biological layer consisting of alluvial mud and various types of microorganisms grows and attaches to the sand media and forms a matrix called schmutzdecke. Changes to several factors, including the quality of raw water, filtration speed, and the addition of media, affect the performance of the slow sand filter unit in producing treated water. Geotextiles can be equipped to improve the performance of a slow sand filter in removing pollutants. The selection of several factors that affect slow sand filter performance can be used as a starting point for the engineering system to determine the best pattern of performance behavior. This approach was carried out by looking at the dynamic behavior patterns of slow sand filter system performance in treating raw water. This research has not yet been conducted extensively. The dynamic behavior pattern approach to the performance of the slow sand filter unit was used to obtain the behavior model for the schmutzdecke layer on the filter. The system dynamic approach focused on treatment scenarios that can determine the behavior of the slow sand filter system. Several factors were assessed, including temperature, turbidity, nutrient concentration, algal concentration, bacteria and dissolved oxygen. Model simulation results show that the comparison of C: N: P values affected the performance of the schmutzdecke layer in removing total coli. The slow sand filter unit was capable of producing treated water with a total amount of coli equal to 0 on the C: N: P values of 85: 5.59: 1.25, respectively, and a 9 cm geotextile thickness.

Behavior of schmutzdecke with varied filtration rates of slow sand filter to remove total coliforms

The previous research showed that slow sand filtration (SSF) can remove the total coli by approximately 99% because of the schmutzecke layer in the filter. The presented study aimed to complete the previous research on SSF, especially on the schmuztdecke layer mechanism, to remove total coli. Total coli is a parameter of water quality standard in Indonesia, and the behavior of schmutzdecke affects the total coli removal. In the present study, the raw water from Amprong River was treated using horizontal roughing filter (HRF) and SSF. The variations in SSF rate used were 0.2 and 0.4 m/h. Total coliforms were analyzed using the most probable number test, and schmutzdecke visualization was conducted through scanning electron microscopy–energy-dispersive X-ray spectroscopy (SEM–EDX). The best coliform concentration in water treated by the combination of HRF and SSF was 4,386 colonies per 100 mL of sample using the filtration rate of 0.2 m/h, and its removal efficiency was 99.60%. However, the quality of water treated by the combination of HRF and SSF did not meet the drinking water quality standard because the removal of total coli must be 100%. The SEM–EDX visualization results in schmutzdecke showed that the average bacteria in the schmutzdecke layer were small, white, opaque, and circular, with entire edge and flat elevation. The Gram test results showed that the schmutzdecke bacteria consisted of Gram-positive and Gram-negative bacteria with basil as the common cell form.

Replicating the microbial community and water quality performance of full-scale slow sand filters in laboratory-scale filters

Previous laboratory-scale studies to characterise the functional microbial ecology of slow sand filters have suffered from methodological limitations that could compromise their relevance to full-scale systems. Therefore, to ascertain if laboratory-scale slow sand filters (L-SSFs) can replicate the microbial community and water quality production of industrially operated full-scale slow sand filters (I-SSFs), eight cylindrical L-SSFs were constructed and were used to treat water from the same source as the I-SSFs. Half of the L-SSFs sand beds were composed of sterilized sand (sterile) from the industrial filters and the other half with sand taken directly from the same industrial filter (non-sterile). All filters were operated for 10 weeks, with the microbial community and water quality parameters sampled and analysed weekly. To characterize the microbial community phyla-specific qPCR assays and 454 pyrosequencing of the 16S rRNA gene were used in conjunction with an array of statistical techniques. The results demonstrate that it is possible to mimic both the water quality production and the structure of the microbial community of full-scale filters in the laboratory - at all levels of taxonomic classification except OTU - thus allowing comparison of LSSF experiments with full-scale units. Further, it was found that the sand type composing the filter bed (non-sterile or sterile), the water quality produced, the age of the filters and the depth of sand samples were all significant factors in explaining observed differences in the structure of the microbial consortia. This study is the first to the authors' knowledge that demonstrates that scaled-down slow sand filters can accurately reproduce the water quality and microbial consortia of full-scale slow sand filters.