Abstract
This study investigates backwashing strategies under full scale conditions with analysis focused on pairing washed filter media associated turbidity with dissolved organic carbon (DOC) removal. A core sampling monitoring program was implemented on full-scale biofilters at two water treatment plants to assess these monitoring tools under varying seasonal and operational conditions. Despite processing the same water source with similar pre-treatment methods (flocculation, coagulation, sedimentation), notable differences were observed between biofilters, showing the importance of site-specific evaluations. The findings indicate that backwashing that included air scouring was more effective than hydraulic backwashing alone in floc removal, with reduced air scour durations maintaining efficiency. Ineffective backwashing led to high floc retention turbidity across filter depths, particularly in systems relying solely on hydraulic wash. Interestingly, filters using hydraulic backwashing, despite failing to meet floc retention guidelines, achieved higher DOC removals suggesting that current floc retention guidelines may be too low for biofilters. Additionally, Fluorescence Excitation-Emission Matrix (FEEM) analysis indicated higher soluble microbial by-product fluorescence in specific filter groups, correlating with higher effluent DOC levels. These findings highlight the distinct nature of biofilters compared to conventional filters, and the importance of additional considerations for backwashing practices and effective DOC removal emphasizing the need for tailored monitoring tools with biofiltration.
Introduction
In the field of water treatment, biofiltration (biofilters) promotes the growth of microorganisms on filter media (e.g. sand, granular activated carbon, anthracite) to aid in the digestion and removal of biodegradable carbon in addition to the standard removal of suspended solids through size exclusion and adhesion. The goal of dissolved organic carbon (DOC) removal is to improve the aesthetic quality and safety of drinking water through the removal of disinfection by-product precursors, colour causing organics, improved water biostability, and reduction in odor causing compounds [16, 23]. The performance of these biofiltration systems, particularly in terms of DOC removal, is often governed by the dynamics within the filter bed, which include the complex interplay of factors like floc retention, filter media choice, empty bed contact time (EBCT), temperature and backwashing strategies [3].
Filtration system design primarily focuses on filter media selection and hydraulic loading rates. However, operational factors such as backwashing strategies, seasonal performance variations, and solids buildup, require ongoing adjustments to setpoints and management strategies. The American Water Works Association (AWWA) provides guidance on maintaining conventional filters [18]. One important aspect of filter monitoring is floc retention analysis, a technique to characterize and quantify adhered material to media [13], that has been applied to conventional filters and pressure sand filters used in drip irrigation, however research with biofilters is lacking [7, 9, 17, 21, 29].
Biofilters differ significantly from conventional filters, warranting the development of dedicated monitoring and maintenance strategies. The critical distinction is the biological component of these filters, which can be sensitive to environmental factors such as temperature and backwashing. Studies have shown that cold temperatures (< 5 °C) can decrease microbial metabolic activity, resulting in slower substrate consumption and reduced biomass development [8, 11, 26, 27]. This reduction in microbial activity may result in a decreased capacity to remove natural organic matter (NOM) and DBP precursors, thereby compromising water quality. In addition, full scale evaluation of seasonal temperature changes with biofilters is still somewhat limited suggesting that additional research is needed to clarify temperature effects.
Backwashing methods are intended to alleviate the challenges associated with suspended solids and floc accumulation in filter beds. Hydraulic backwashing refers to the upflow of water only through the media bed while air scour introduces air into the filter bed to agitate the media and dislodge trapped particles and is used in conjunction with hydraulic backwashing [22, 24]. Ahmad and Amirtharajah, [1] examined the forces needed to remove biological and non-biological particles from filter beds, at pilot scale, during backwashing. Their findings suggested that biofilters require stronger shear forces to dislodge biological particles compared to particulate matter alone.
Ramsay et al. [22] explored the role of air scouring on pilot filter columns and demonstrated that grain displacement occurred during air scour at the surface of the media, known as sand boiling, and that longer durations were not efficient at improving particulate removal. Piche et al. [20] compared backwashing strategies for biofilters and conventional filters, concluding that air scouring minimized head loss development for up to 72 h post-backwash. Ikhlef and Basu, [11] observed that backwashing using air scouring techniques improved DOC removal in lab-scale biofilters compared to hydraulic backwashing within a 20–30% bed hydraulic bed expansion, beyond this DOC removal decreased, suggesting that a balance may exist between “sufficient” and “excessive” shear forces on biological filters that are determinantal to DOC removal. While optimal bed expansion for conventional filters have been reported between 15 and 50% [6, 11, 18, 28], the ideal range for biofilters remains unclear. Further research is needed to determine how bed expansion influences backwashing efficiency in biofilters, and the development of improved backwash efficiency indicators could help optimize operational parameters while minimizing energy and water use.
Biological processes such as dissolved organic carbon (DOC) removal, nitrification, and ammonia oxidation can be reduced when chlorinated water is used for backwashing in biofilters, as its oxidative properties negatively affect biomass stability [15, 27, 30, 31]. Meanwhile Liu et al. [14] reported that on a bench-scale level, chlorinated backwash water did not impact BOM removal levels at temperatures of 5 °C and 20 °C. Given the varying results across studies, further research is needed to validate these findings under real-world operational conditions. Most investigations have been conducted at the pilot or bench scale, and a lack of full-scale studies limits the ability to optimize backwashing strategies for different biofilter applications. Additionally, for filters that lack air scouring capacity, alternative methods must be explored to improve backwash efficiency without compromising biofilter function.
This research explores the comparative effectiveness of various backwashing methods in maintaining full-scale filter bed performance. Two sets of biofilters at two water treatment plants over a period of two years were evaluated through conducting a core sampling program in combination with DOC analysis. By examining the role of different backwashing strategies, we aim to uncover how backwashing strategies influence floc retention and the removal of dissolved organic matter in biofilters. The results of this investigation will contribute to the development of optimized backwashing protocols, ensuring more efficient and sustainable biofiltration processes in drinking water treatment.
Methods and materials
This research was conducted in Ottawa, Ontario, (Canada) at two water treatment plants Lemieux Island (Plant A) and Britannia (Plant B) Water Purification Plant treating water from the Ottawa River.
Biofilter design and operation
Each plant operates a total of eighteen dual-media anthracite-sand biofilters, each with a surface area of 135 m2. The upstream treatment processes include coagulation, flocculation, and sedimentation, while disinfection occurs downstream of the filter units. A selected subset of filters from each plant was evaluated: filters A1 (filters 1–12 at Plant A) and filters B1 (filters 13–18 at Plant B), as illustrated in Fig. 1.
Fig. 1.
Cross-sectional view of filter media composition [12]
The anthracite in B1 filters has an effective size (ES) of 0.60 mm compared to 1.06 mm in A1 filters. The sizing distinction is related to the age of the media with B1 anthracite approx. 35 years old and A1 anthracite approx. 15 years old. Similarly, the sand media in B1 filters has an ES of 0.43 mm, while A1 filters maintain a larger ES of 0.53 mm. Detailed sieve analysis data is available in Supplementary Data A. Filters maintain an empty bed contact time (EBCT) between 13 and 24 min, depending on plant production needs. Filters A1 are equipped with stainless steel underdrains, whereas B1 has a gravel support bed with clay underdrains.
Study phases and sampling procedures
Core sampling was completed to extract filter media from the filter beds using a 2-inch diameter PVC sampling probe. Samples were collected both at the end of a filter run (before backwashing) and at the start of a filter run (after backwashing), seen in Fig. 2. Media was retrieved from four locations and at six depth levels (0–5 cm, 5–15 cm, 15–30 cm, 30–45 cm, 45–60 cm, and 60–75 cm). Samples were collected from the center of the filter bed, at two adjacent sites with approx. 50 g of media collected at each location and depth in the filter bed. In some cases, corner samples were collected as well. Sample material was collected and homogenized for analysis – with 6–12 replicates run from each sample collection and depth.
Fig. 2.
Core sampling procedure [12]
The study was conducted in six sampling periods between July 2023 and January 2025. Phase 1 involved observational data collection across three temperature ranges, summer (> 15 °C), fall (5–15 °C) and winter (0–5 °C), denoted by 1-S, 1-F and 1-W. Phase 2 examined the effects of modified backwash procedures and included assessing changes under the three temperature conditions (2-S, 2-F and 2-W). The B1 filters undergo backwashing through a multi-step process, beginning with a surface wash using sweep surface agitators, followed by a hydraulic wash, and concluding with an extended terminal subfluidization wash (ESTW). In comparison, A1 filters utilize an air scour, followed by a hydraulic wash, and finish with an ESTW. Figure 3 summarizes the different backwashes used over the different sampling periods.
Fig. 3.
Backwashing procedures for A1 and B1 filters by experimental phase
Analytical methods
Floc retention analysis
Floc retention analyses were performed on the collected filter media following the procedure outlined in Fig. 4 The condition of the filter media was assessed based on the guidelines provided in Table 1.
Fig. 4.
Floc retention analyss step-by-step procedure (adapted from [19])
Table 1.
Floc retention turbidity and associated filter media condition for conventional filters [19]
| Floc retention turbidity | Filter media condition (Conventional Filters) |
|---|---|
| 0–30 NTU | Clean—unripened filter (expect long periods of turbidity breakthrough) |
| 30–60 NTU | Clean—ripened filter (filter performance optimal) |
| 60–120 NTU | Slightly dirty (sub-optimal, but should perform satisfactorily) |
| > 120 NTU | Dirty (re-evaluate the backwashing procedure/system) |
| > 300 NTU | Mudball problems |
DOC and feem analysis
Water samples were collected from both the influent and effluent of the filters at specified time intervals: before backwash (end of a filter run), 5 min after backwash, 60 min after backwash, and 90 min after backwash. The samples were collected in 40 mL amber glass bottles, immediately filtered through 0.45 μm filter paper, and then transferred into clean 40 mL clear glass bottles, which were wrapped in aluminum foil to protect from light.
Filter media samples were collected from the decanted floc retention analysis water sample. The wash water from this analysis was collected directly from a 500 mL flask, filtered through 0.45 μm filter paper, and transferred into clean 40 mL clear glass bottles wrapped in aluminum foil.
All samples were preserved with 2 mL of 4% phosphoric acid to inhibit bacterial activity and prevent DOC reduction [2]. Samples were stored at 4 °C.
Samples underwent both DOC analysis and FEEM analysis. Dissolved organic carbon (DOC) analysis was conducted using a Shimadzu TOC-VCPH/CPN analyzer, utilizing the combustion oxidation method outlined in Standard Methods 5310 C [2]. FEEM analysis was performed using a Horiba Aqualog® spectrometer. The samples underwent both 3D EEM (Excitation Emission Matrix) and absorbance scans. The scanning parameters included an integration time of 2 s, medium CCD gain, a 5 nm step increment, and an excitation wavelength range from 200 to 600 nm. Corrections applied to each scan included inner filter effect correction, 1st and 2nd order Rayleigh masking (at 10 nm), and normalization based on the blank scan of a distilled water sample.
Results and discussion
Effectiveness of backwash strategies
Floc retention turbidity data was collected over all study phases (refer to Figs. 6 and 7) with a goal of assessing if backwash regime and season influenced floc retention turbidity results. Phase 1 was a monitoring phase where parameters were analysed against standard seasonal operations at the respective plants.
Fig. 6.
A1 filters’ floc retention turbidity of filter media before and after backwash during each study phase. Dotted red line indicating the 60 NTU clean media threshold
Fig. 7.
B1 filters’ floc retention turbidity of filter media before and after backwash during each study phase. Dotted red line indicating the 60 NTU clean media threshold
In the A1 filters, floc retention turbidity before backwashing is highest in the upper layers, corresponding to the anthracite media, while the sand layers (60 cm and 75 cm) exhibit lower turbidity. After backwashing, turbidity decreases to below 60 NTU, aligning with the conventional clean media threshold guideline.
In contrast, the B1 filters exhibit consistently higher floc retention turbidity over the entire bed depth before backwashing, with no distinction between the anthracite and sand layers. After backwashing, turbidity remains unchanged, indicating that the process was ineffective in removing trapped particles. Furthermore, physical examination of the filter bed revealed surface layer cracking in the anthracite and the presence of small mudballs in core samples (albeit they were not consistently found throughout the study) (Fig. 5). These findings align with the consistently high floc retention turbidity observed throughout the study phases.
Fig. 5.
B1 filters a surface cracking of top anthracite layer b mudballs found in 30 cm depth
These results provide a comparison between hydraulic backwashing (surface wash + water wash) and air scouring methods. The hydraulic wash used on the B1 filters proved ineffective in reducing floc retention turbidity within the filter bed. Throughout the study phases, the backwash velocity was increased from 45 m/h to 54 m/h, reaching the maximum capacity of the backwash pumps. Literature suggests backwash velocity ranges from 37.5 m/h to 50 m/h [13]. As an additional measure, backwash duration was increased from 5 min to 7 min. However, despite increases in both velocity and duration, no improvement in floc retention was observed. The size of the anthracite in the B1 filters has an ES of 0.60 mm, which falls outside recommended filter media specifications (0.8–2 mm) [5]. It is suspected that these filters may be running sub-optimally due to the condition of the media with respect to the floc retention results. The changes in backwashing procedures over the experimental phases, were not successful in reducing the build up of floc within the filter bed, suggesting a larger issue with filter health. As the average media size is now less than recommended, which is likely associate with the high media age, it may be at end of its operational lifespan or may need additional chemical washing to rehabilitate. Additional research that explores media properties such as surface roughness, etc. maybe useful to determine additional impacts of media age on the filter bed operations.
The A1 filters, which are run at shorter and lower backwash velocities than B1 filters (refer to Fig. 3), were able to remove the floc buildup over a filter run (< 60 NTU). For instance, post backwash A1 filter average floc retention turbidity at 5 cm and 75 cm in Summer (1-S) were 76.3 NTU and 64.5 NTU; whereas B1 filter values were 239 NTU and 144 NTU, respectively; with differences by depth statistically significant p < 0.05. Results for Fall (1-F) showed similar differences (p < 0.05) where again post backwash A1 floc retention turbidity values of 117 NTU and 115.6 NTU at 5 and 75 cm respectively, and post backwash B1 209 NTU and 210 NTU, respectively. This trend continued in the winter season as well; the overall global average across all depths for post backwash A1 versus B1 filters were 56 ± 29 NTU and 200 ± 54 NTU. The primary distinction in backwashing design—air scouring—was observed to be the key difference in the filter operations pointing to the enhanced removal of accumulated floc throughout the study phases. The preliminary air scouring step allowed for a shorter and less abrasive water wash, which remained effective. To further evaluate the system’s efficiency and testing for potential energy savings, the air scour duration was reduced from 10 min to 5 min. Despite these reduced backwash conditions, the process remained effective. It is important to note that the A1 filters are closer to recommended filter media size (1.03 mm) for anthracite which is believed to contribute to better performance with respect to floc retention observed in the filters (Figs. 6 and 7).
Impact of chlorination on floc retention
A1 filters were subject to a period of chlorinated backwashing (CBW) in Phase 1-S of the study. In order to facilitate biological activity in the filters, plants generally operate with non-chlorinated backwash (NCBW) [19]; however, at times full-scale operation requires plant changes that necessitate a chlorinated water (0.87–1.52 mg/L free chlorine) being utilized within the filter system [4]. The A1 filters experienced: a 30-day period in Phase 1-S (July 2023) due to operational and maintenance activities at the plant. Figure 8 summarizes the floc retention results from phase 1-S comparing samples taken before, after 30 days of chlorinated backwashing, and post-transition to monitor recovery.
Fig. 8.
Floc retention results after 30 day backwash under chlorinated (CBW) versus non-chlorinated (NCBW) backwashing conditions throughout phase 1-S (Summer 2023) for A1 Filters
Prior to the transition to CBW, floc retention in the filter reached a maximum of 112 ± 26 NTU at the 30 cm level. Across all depths, the mean floc retention values ranged from 63 NTU to 73 NTU, while standard deviations varied considerably, from 5 to 26 NTU. Notably, the upper layers of the filter exhibited greater variability, indicating more fluctuation in floc retention compared to the deeper layers. After 30 days of CBW (July 2023), the maximum turbidity increased to 164 NTU at the 5 cm depth, marking a substantial rise. Additionally, the upper layers of the filters exhibited greater variability with standard deviations 38 and 27 NTU, at the 5 cm and 15 cm depths, respectively. These findings suggest that chlorinated backwashing may have a more pronounced impact on floc stability in the upper filter layers compared to deeper sections. Two months after reverting to non-chlorinated backwashing, samples taken in September 2023 indicate that the filters continued to exhibit elevated floc retention turbidity values and high variability, suggesting that recovery to pre-CBW conditions (June 2023) had not yet been achieved. This prolonged impact highlights the lasting effects of chlorinated backwashing on filter stability. However, by November 2023, when water temperatures ranged from 5 to 15 °C, floc retention turbidity levels recovered, averaging below 60 NTU, suggesting a gradual restoration of filter performance.
Dissolved organic carbon
Filter media DOC
The floc retention analysis conducted on the media produces a water sample that is tested for turbidity. This sample, which contains floc removed from the filter media during the manual shaking process, was filtered and subsequently analyzed for dissolved organic carbon (DOC), as shown in Figs. 9 and 10. This novel approach is intended to investigate if floc retention DOC values may correlate to influent/effluent DOC removal as an indicator of potential biomass retained on the media. The DOC concentration was highest in the B1 filters compared to the A1 filters. Within the B1 filters before backwashing, the highest concentration was at the top of the filter (15 cm), 16.55 ± 0.34 mg/L, and in the winter (December)—his is consistent with the floc retention results, which peak at 341 NTU, in December. However, after backwash, an inverse relationship emerges, with DOC concentrations peaking at the bottom of the filter (75 cm), 16.06 ± 1.24 mg/L, most evidently in the winter sampling. This trend was less pronounced in summer and fall, where DOC levels remained consistent throughout the filter bed. While floc retention turbidity remains uniform along the filter depth, the migration of organic carbon-rich media to the bottom suggests a potential link to mudball formation. As floc-covered media becomes denser, it settles at the bottom during post-backwash resettlement, contributing to this phenomenon.
Fig. 9.
B1 filters filter media DOC results
Fig. 10.
A1 filters filter media DOC results
Comparatively, the A1 filters maintained low DOC levels, with the highest concentration (13.82 ± 0.14 mg/L) observed at the top of the filter in August, gradually decreasing throughout the study period. This decline coincided with the implementation of chlorinated backwashing, which may have contributed to the reduction in DOC levels from August to November. Following backwashing, DOC concentrations remained consistently low throughout the filter bed depth, aligning with the observed low floc retention turbidity.
A subset of the filtered floc retention DOC samples were also tested for FEEM. The results provide deeper insight into the types of DOC present (Fig. 11). Three distinct regions of excitation and emission wavelengths are seen at roughly Ex/Em 250/420 nm, Ex/Em 325/400 nm, and Ex/Em 275/300 nm, which can be classified as fulvic-like, humic-like, and soluble microbial by-product (SMP)-like substances, respectively [10, 25]. The FEEM results are consistent with the floc retention turbidity findings, which demonstrate more efficient backwashing in A1 filters compared to B1 filters, as evidenced by the reduction in the DOC fingerprint before and after backwashing. The SMP region in B1 filters remains constant before and after backwashing, whereas in A1 filters, it fluctuates. This difference may be linked to the chlorinated backwashing process, which could be impacting biomass activity post-backwash.
Fig. 11.
FEEM results of filter media DOC samples at 30 cm in Phase 2-F for a B1 filter before backwash b B1 filter after backwash c A1 filter before backwash d A1 filter after backwash
Influent and effluent DOC
Biofilters, without ozonation, are expected to remove on average 14% of DOC [27]; in this research DOC removal was examined across seasons as an indicator of biofilter health. The results (Tables 2 and 3), indicate DOC removal varied significantly across temperature ranges (p < 0.05). The following sections present a comparative analysis of B1 vs. A1 filters and their relationship to floc retention, biomass activity, and DOC removal efficiency.
Table 2.
Influent and effluent DOC concentrations for B1 filters taken over study phase 2-S, 2-F and 2-W
| Phase | Influent/Effluent | Average DOC Conc. (mg/L) |
|---|---|---|
| 2-S | Influent | 3.18 ± 0.76 |
| 2-S | Effluent | 2.64 ± 0.37 |
| Average DOC Removal (%) | 16.92 ± 1.11% | |
| 2-F | Influent | 3.36 ± 0.47 |
| 2-F | Effluent | 2.90 ± 0.57 |
| Average DOC Removal (%) | 13.81 ± 8.10% | |
| 2-W | Influent | 4.69 ± 0.66 |
| 2-W | Effluent | 4.75 ± 0.31 |
| Average DOC Removal (%) | − 1.37 ± 5.71% | |
Table 3.
Influent and effluent DOC concentrations for A1 filters taken over study phase 2-F and 2-W
| Phase | Influent/Effluent | Average DOC Conc. (mg/L) |
|---|---|---|
| 2-F | Influent | 3.29 ± 0.59 |
| 2-F | Effluent | 3.47 ± 0.61 |
| Average DOC Removal (%) | 5.68 ± 2.46% | |
| 2-W | Influent | 4.83 ± 0.81 |
| 2-W | Effluent | 4.68 ± 0.69 |
| Average DOC Removal (%) | 3.24 ± 2.90% | |
For the B1 filters (Table 2), the highest average DOC removal was observed during the summer months (16.92 ± 1.11%), whereas removal dropped to near zero (− 1.37 ± 5.71%) in winter (< 5 °C, notable in Ottawa, water temperatures dropped to 0.4 °C during the coldest months of January and February). This seasonal variation suggests that colder temperatures impaired the biological treatment process. These findings align with previous studies demonstrating reduced biological activity and DOC removal efficiency under cold conditions [3, 27]. In A1 filters (Table 3) where floc retention values met expected guidelines values (60 NTU), DOC removal was only 3.24 ± 2.90%. However, it is important to note, that the peak average removal is based on fall conditions as access to influent and effluent sample ports was not available during the summer season due to full-scale plant maintenance activities. These results indicate a potential conflict between guideline floc retention values and a healthy biofilter which aims to reduce DOC content. Adhering rigidly to conventional filtration guidelines for floc retention turbidity (< 60 NTU) may unintentionally limit biofilter performance. These findings are further supported by additional ATP analysis conducted during the study period. The B1 filters, which retained more floc, exhibited higher ATP levels (86.96–176.49 ng ATP/g) compared to the A1 filters (27.81–40.79 ng ATP/g). This points to a reduced biomass concentration, lower biological activity and consequently, lower DOC removals.
Both filters consistently met guideline criteria of less than 100 ug/L THM and less than 80 ug/L HAA, as seen in Table 4. Overall, Plant B demonstrated slightly better removal of DBPs compared to Plant A. This may be connected to the biomass data and floc retention turbidity, where better DOC removal is indicative of better biological treatment of the water and better biofiltration; however, it is also important to note that these may be further attributed to the slight operational differences at the plant as well. Thus, this continues to support that the floc retention guidelines are too strict for biofilters.
Table 4.
Monthly DBP sampling of treated water at Plant A and Plant B
| Study Phase | Sample Month | THM (µg/L) | HAA5 (µg/L) | ||
|---|---|---|---|---|---|
| Plant A | Plant B | Plant A | Plant B | ||
| 2-S | June | NA | 38.5 | NA | 29.4 |
| July | NA | 47.6 | NA | 32 | |
| August | NA | 48.4 | NA | 29.8 | |
| September | NA | 36 | NA | 25.7 | |
| 2-F | October | 52 | 35 | 47.3 | 34.2 |
| November | 38.8 | 28.8 | 29.2 | 22.2 | |
| 2-W | December | 30.8 | 25.2 | 31.3 | 21.6 |
Figure 12 shows the FEEM results for the effluent samples taken 1-hour post-backwash (a and b) and at the end of the filter run (b and d) for Plant A and B. The fluorescence patterns observed in these effluent samples are consistent with those seen in the filter media FEEM results (Fig. 11), suggesting that the same three organic fractions persist in the effluent. Higher fluorescence intensities in the Plant A filters are seen, particularly within the Ex/Em 275/300 nm region (SMP-like substances). This is consistent with the DOC concentration results which are higher at Plant A than Plant B, suggesting differences in effluent water quality from the filter groups.
Fig. 12.
FEEM Results of effluent DOC samples in phase 2-F for a B1 filter 1 h into Filter Run b B1 filter at End of Filter Run c A1 filter 1 h into Filter Run d A1 filter at End of Filter Run
Conclusions
The implementation of a long-term monitoring program enabled a comprehensive analysis of full-scale biofilters under varying seasonal and operational conditions. This research provides valuable insights into how key parameters, particularly backwashing techniques, influence biofiltration performance in terms of floc retention turbidity and DOC removal. Despite processing the same water source with similar pre-treatment methods, significant differences were observed between filter groups at Plant A (hydraulic backwashing with air scour) and Plant B (hydraulic backwashing only). These findings highlight the importance of continuous evaluation of full-scale filters as part of routine maintenance practices. The following conclusions were drawn from the study:
The amount of floc retained on the filter media post-backwash varied significantly between filter groups using hydraulic backwashing and those utilizing air scouring with global averages of 200 NTU vs. 59 NTU, respectively. Air scouring was the more effective method for floc removal, and reducing the air scour duration from 10 min to 5 min did not compromise efficiency and can contribute to decreased energy usage for maintaining the filters.
The biofilters utilizing hydraulic backwashing achieved greater DOC removals (up to 16%) and demonstrated higher ATP levels (86–176 ng ATP/g) than those with air scour and that met floc retention guidelines with DOC removal (up to 3%) and lower ATP levels (27–41 ng/ATP/g). This suggests that the current floc retention threshold < 60 NTU may be too low for biofilters and that < 120 NTU maybe more appropriate with additional research recommended. Overly aggressive backwashing may hinder microbial growth, potentially compromising the biofiltration process and its effectiveness in DOC removal.
Chlorinated backwashing had a lasting impact on floc retention stability, particularly in the upper filter layers. Floc retention turbidity increased, from 112 NTU to 164 NTU with higher variability observed at shallower depths. Two months after transitioning back to non-chlorinated backwashing, pre-chlorinated backwashing conditions had not been fully restored. However, four months post-CBW, coinciding with a seasonal temperature decrease, floc retention turbidity stabilized below 60 NTU, indicating a gradual recovery.
Fluorescence Excitation-Emission Matrix (FEEM) analysis revealed consistent organic fractions in both filter groups with A1 filters showing higher fluorescence intensity in soluble microbial by-product (SMP)-like regions, correlating with increased effluent DOC concentrations.
This research strengthens the call for additional development of biofilter centric operational guidelines and monitoring practices. Future research should focus on optimizing backwashing strategies, assessing long-term floc stability, and developing biofilter-specific performance metrics to enhance their efficiency and reliability.
Acknowledgements
The authors gratefully acknowledge the financial support provided by the City of Ottawa, Carleton University, the Natural Sciences and Engineering Research Council of Canada (NSERC), the CREATE OPSIDIAN NSERC program, and the Mitacs Accelerate Program. The authors also thank Issabella Anim who assisted with filter sampling.
Author contributions
ZJ: compiled and analyzed data, wrote first draft of paper, edited paper, performed core sampling of filters, conducted laboratory tests required for data collection JR: coordinated logistics for filter sampling and provided feedback on results and contributed to reviewing paper drafts JE: reviewed and provided feedback on results and contributed to reviewing paper drafts OB: Thesis supervisor; designed the experimental framework, provided feedback on results, and contributed to the writing process.
Data availability
All data supporting the findings of this study are available within the paper and/or supporting documentation.
Declarations
Ethics approval and consent to participate
Not applicable.
Competing interests
The authors declare no competing interests.
References
- 1.Ahmad R, Amirtharajah A. Detachment of particles during biofilter backwashing. J Am Water Works Assoc. 1998;90(12):74–85. [Google Scholar]
- 2.American Public Health Association, American Water Works Association, & Water Environment Federation. In: Baird RB, Eaton AD, Rice EW, editors. Standard Methods for the Examination of Water and Wastewater. 24th ed. American Water Works Association; 2023.
- 3.Basu OD, Dhawan S, Black K. Applications of biofiltration in drinking water treatment—A review. J Chem Technol Biotechnol. 2015;91(3):585–95. 10.1002/jctb.4860. [DOI] [Google Scholar]
- 4.City of Ottawa. (2023). Lemieux Purification Plant Operator’s Process Spreadsheet [Dataset].
- 5.Crittendon JC, Trussel RR, Howe KJ, Tchobanogloud G. In: Sons, editor. MHW’s Water Treatment Principles and Design. John Wiley; 2012. 10.1002/9781118131473.ch11. [DOI]
- 6.De Souza FH, Pizzolatti BS, Sens ML. Backwash as a simple operational alternative for small-scale slow sand filters: From conception to the current state of the art. J Water Process Eng. 2021;40:101864. [Google Scholar]
- 7.Duran-Ros M, Pujol J, Pujol T, Cufí S, Arbat G, De Ramírez F, Puig-Bargués J. Solid Removal across the Bed Depth in Media Filters for Drip Irrigation Systems. Agriculture. 2023;13(2):458. 10.3390/agriculture13020458. [DOI] [Google Scholar]
- 8.Evans AN, Carter JT, Kuhnel B, Nystrom VE. Seasonal impacts on biofiltration acclimation dynamics and performance. AWWA Water Sci. 2021;3(3):e1225. 10.1002/aws2.1225. [DOI] [Google Scholar]
- 9.Haarhoff J, Van Staden S. Technical note: Measurement and expression of granular filter cleanliness. Water SA. 2013;39(5):701. 10.4314/wsa.v39i5.15. [DOI] [Google Scholar]
- 10.Hidayah EN, Pachwarya RB, Cahyonugroho OH, Ramanathan AL. Characterization of Molecular Weight–Based Fluorescent Organic Matter and Its Removal in Combination of Constructed Wetland with Activated Sludge Process. Water Air Soil Pollut. 2020;231(2):41. 10.1007/s11270-020-4405-5. [DOI] [Google Scholar]
- 11.Ikhlef S, Basu OD. Influence of backwash regime on biofilter performance in drinking water treatment. J Chem Technol Biotechnol. 2017;92(7):1777–84. 10.1002/jctb.5178. [DOI] [Google Scholar]
- 12.Jeaurond Z, Campbell A, Elliott J, Rode J, Basu OD. (2025). Does Your Backwash Matter? Influences on Filter Media Health. In: Ng, K.T.W., Basu, O., Robinson, C, editors Proceedings of the Canadian Society for Civil Engineering Annual Conference 2024, Volume 5. CSCE 2024. Lecture Notes in Civil Engineering, vol 695. Springer, Cham. 10.1007/978-3-031-97689-6_16 [DOI]
- 13.Kawamura S. Integrated Design and Operation of Water Treatment Facilities. John Wiley & Sons Inc; 2000.
- 14.Liu X, Huck PM, Slawson RM. Factors Affecting Drinking Water Biofiltration. J Am Water Works Assoc. 2001;93(2):90–101. [Google Scholar]
- 15.McKie MJ, Taylor-Edmonds L, Andrews SA, Andrews RC. Engineered biofiltration for the removal of disinfection by-product precursors and genotoxicity. Water Res. 2015;81:196–207. 10.1016/j.watres.2015.05.034. [DOI] [PubMed] [Google Scholar]
- 16.McKie MJ, Curl JM. Lessons Learned in Biofiltration. J Am Water Works Association. 2023;115:30–8. [Google Scholar]
- 17.Mesquita M, De Deus FP, Testezlaf R, Diotto AV. Removal efficiency of pressurized sand filters during the filtration process. Desalin Water Treat. 2019;161:132–43. 10.5004/dwt.2019.24285. [DOI] [Google Scholar]
- 18.Nix DK, Taylor JS. Chapter 8—Backwash Bed Expansion Analysis. Filter Evaluation Procedures for Granular Media. 2nd ed. American Water Works Association; 2018a. pp. 61–74.
- 19.Nix DK, Taylor JS. Chapter 24—Biofilters. Filter Evaluation Procedures for Granular Media. 2nd ed. American Water Works Association; 2018b. pp. 229–33.
- 20.Piche A, Campbell A, Douglas I, Basu OD. Analyzing the performance of biological versus conventional drinking water filtration under warm and cold water conditions: a pilot scale study. Chemosphere. 2019;234:52–61. 10.1016/j.chemosphere.2019.05.180. [DOI] [PubMed] [Google Scholar]
- 21.Pujol T, Duran-Ros M, Arbat G, Cufí S, Pujol J, De Ramírez F, Puig-Bargués J. Pressurised sand bed filtration model: Set up and energy requirements for a filtration cycle. Biosyst Eng. 2024;238:62–77. 10.1016/j.biosystemseng.2024.01.001. [DOI] [Google Scholar]
- 22.Ramsay L, Du F, Lund M, He H, Søborg DA. Grain displacement during backwash of drinking water filters. Water Supply. 2021;21(1):356–67. 10.2166/ws.2020.300. [DOI] [Google Scholar]
- 23.Sadiq R, Rodriguez MJ. Disinfection by-products (DBPs) in drinking water and predictive models for their occurrence: a review. Sci Total Environ. 2004;321(1–3):21–46. 10.1016/j.scitotenv.2003.05.001. [DOI] [PubMed] [Google Scholar]
- 24.Slavik I, Jehmlich A, Uhl W. Impact of backwashing procedures on deep bed filtration productivity in drinking water treatment. Water Res. 2013;47(16):6348–57. 10.1016/j.watres.2013.08.009. [DOI] [PubMed] [Google Scholar]
- 25.Sriboonnak S, Induvesa P, Wattanachira S, Rakruam P, Siyasukh A, Pumas C, Wongrueng A, Khan E. Trihalomethanes in water supply system and water distribution networks. Int J Environ Res Public Health. 2021;18(17):9066. 10.3390/ijerph18179066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Stoddart AK, Gagnon GA. Full-scale prechlorine removal: impact on filter performance and water quality. J Am Water Works Assoc. 2015;107(12):638–47. 10.2307/jamewatworass.107.12.e638. [DOI] [Google Scholar]
- 27.Terry LG, Summers RS. Biodegradable organic matter and rapid-rate biofilter performance: a review. Water Res. 2018;128:234–45. 10.1016/j.watres.2017.09.048. [DOI] [PubMed] [Google Scholar]
- 28.Turan M. Backwashing of granular media filters and membranes for water treatment: a review. AQUA Water Infrastruct Ecosyst Soc. 2023;72(3):274–98. 10.2166/aqua.2023.207. [DOI] [Google Scholar]
- 29.Van Staden S, Haarhoff J. The use of filter media to determine filter cleanliness. Phys Chem Earth Parts A/B/C. 2011;36(14–15):1135–40. 10.1016/j.pce.2011.07.067. [DOI] [Google Scholar]
- 30.Vázquez BG, Choubert JM, Canler JP, Paul E. Solid build-up and consolidation within biofiltration of wastewater: a new multicriteria analysis to early detect irreversible clogging. J Water Process Eng. 2024;65:105796. [Google Scholar]
- 31.Wert EC, Neemann JJ, Rexing DJ, Zegers RE. Biofiltration for removal of BOM and residual ammonia following control of bromate formation. Water Res. 2008;42(1–2):372–8. 10.1016/j.watres.2007.07.028. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data supporting the findings of this study are available within the paper and/or supporting documentation.