The study was carried out at Rohingya camps in Cox’s Bazar, Bangladesh. Details locations of the sampling sites and points are depicted in Figure 1. The focus of this study was on the FSM technologies of the Rohingya camps, and the site selection was completed based on the location of five different FSM technologies in different camps under different NGOs and WASH partners of UNICEF. The selected technologies were UPF, ABR, LSP, WSP and DEWATS. Schematic diagrams of the technologies are depicted in (Supplementary Figures S1–S5). Samples were collected from 17 different plants (4 from each of ABR and UPF, three from each of LSP, DEWATS and WSP) of five different technologies throughout the year from both the inlet (raw sludge) and the outlet (treated sludge) of the plants. Treatment plants were symbolized from P1 through P17. Details information on the FS treatment technologies and plants along with basic structural characteristics based on the data collected by instant interviewing the respective responsible personnel of the plants are provided in Table 1. This study was carried out following a 12-round of sampling for 12 months to cover seasonal variation, so the total number of samples was 408 (17 × 2 × 12).

The samples were collected from December 2019 to April 2021. The composite sampling method was followed for collecting samples from both the inlet and outlet of individual treatment plant. Inlet samples were collected from the vacuum-truck during pouring into the inlet chamber at different time intervals-initial, midpoint and last stage. For the outlets, in the case of LSP and WSP, samples were collected from different corners and depth of the polishing ponds, and mixed in a large container for the pupose of composite sampling. For rest of the plants, outlet samples were collected from the discharge pipeline at different time intervals. The samples were taken into high-density polyethylene (HDPE) plastic bottles (Nalgene, Rochester, USA). For microbiological analysis, samples were collected in pre-sterilized bottles. In the case of total nitrogen (TN), and COD, samples were collected in pre-acidified acid-washed bottles. Samples for testing BOD, TS, TSS and phosphate (PO₄) were collected in sterile HDPE bottles. The pH and conductivity samples were measured onsite via a portable pH/conductivity meter (Hanna Instruments, HI98130, USA). To ensure data accuracy and representativeness, duplicate samples were collected from both the inlet and outlet of each treatment plant at each sampling round. In addition, a field blank was also included in each sampling round to monitor potential contamination during sample collection. After collection, the samples were kept in a cool box with a sufficient amount of ice packs to maintain a temperature between 4°C and 10°C. Then, the samples were transported to the Laboratory of Environmental Health, icddr,b, Dhaka via air shipment. Analysis of physicochemical and microbiological parameters was done within 24 h of the collection of samples.

Data analysis was performed using Statistical Programming Language R (version 4.1.2) and Microsoft Excel 2019 (R Core Team, 2021; Microsoft Corporation, 2019). Due to the skewed distribution of the parameters, their summary statistics among different FS treatment technologies (inlet/outlet) were expressed in terms of the minimum, maximum, median, and interquartile range (IQR) (Schmid and Huber, 2014). To compare the efficiency of the technologies, the removal efficiency in percentage was calculated using the following formula (Otaka et al., 2019) presented in the bar diagram. %   R e m o v a l   E f f i c i e n c y = i n l e t − o u t l e t i n l e t x 100

In addition to this, a bar diagram was also used to present the percentage of plants in a respective technology where the parameters were within the guidelines. Violin plots were used to illustrate the influent and effluent characteristics across different parameters, providing a clear depiction of data distribution and dispersion. Finally, the correlations between the parameters for different treatment technologies in the case of both inlet and outlet were calculated using the Spearman rank correlation method which is suitable for skewed data. Here, only the significant (at a 5% level of significance) correlation coefficients are presented in a correlation plot with a color gradient (deep blue-highly positive, deep red-highly negative). A scale (Table 2), applied to both positive and negative relationships was used to interpret the strength and direction of the relationship shown by the values inside the squares (Evans, 1996).

The outlet characteristics were compared with the Bangladesh Department of Environment (DoE) standards for sewage discharge to the receiving water bodies and land. In the case of some parameters, Bangladesh does not have standard guidelines, therefore, the parameters were compared with EU, FAO and WHO outlet discharge guidelines. For pH, BOD, COD, TSS, E. coli and phosphate, the Environment Conservation Rules 2023 guideline was followed (The Environment Conservation Rules, 2023, 2023). For TN, the outlet values were compared with the European Union discharge guidelines (Lu et al., 2016). And lastly, for helminth and E. coli the WHO guidelines for the safe use of wastewater excreta and greywater were used (World Health Organization, 2006). Standard disccharge guideline of different parameters that were used to compare the current results from outlet (effluent) sample is provided in Table 3.

The results of various physicochemical and microbiological parameters for the inlets and outlets of the selected ABR plants are presented in Supplementary Figure S6 (plants P1, P2, P7, and P16). In the ABR system, E. coli counts in the inlet consistently exceeded 20,000 cfu/100 mL across all plants. For the outlet, none of the plants met the discharge guidelines for E. coli throughout the year, except for P16 in round 5, P7 in rounds three and 5, P2 in round 11, and P1 in rounds 2 and 5. No significant trend was observed in the helminth counts for the inlets of all four ABR plants, which consistently showed high counts. However, outlet helminth counts were within the guidelines for P1 (round 3), P2 (rounds 2, 7, 10, and 11), P7 (rounds 2, 8, and 9), and P16 (rounds 3-7 and 9–12). Notably, P16 exhibited significantly lower helminth counts in all rounds except round 8. Regarding physicochemical parameters, the inlet concentrations of COD and BOD exceeded 400 mg/L and 145 mg/L, respectively, across all plants. Outlet values were within guidelines only for P7 in round 5, P1 for COD in round 5, and P2 for COD in round 5. Total nitrogen levels consistently exceeded 400 mg/L and failed to meet the guidelines. TSS levels were below the guideline for P1 (round 6), P2 (rounds 1 and 5), P7 (round 5), and P16 (rounds 5–7). Phosphate levels were within guideline limits for P1 (rounds 2–4), P2 (rounds 1–5), P7 (rounds three and 5), and P16 (round 5).

Data from DEWATS plants (P10, P11, P12, Supplementary Figure S10) showed no clear trends for inlet and outlet parameters. E. coli consistently surpassed 20,000 cfu/100 mL at the inlet. The outlet adhered to E. coli guidelines in a few plants and rounds (Plant P10: rounds 4 and 7; Plant P11: rounds 1–10; Plant P12: rounds 3 and 11). Helminth counts exceeded the guideline in P12 consistently, while P10 met the guideline in some rounds (rounds 2, 8–12) and P11 in others (rounds 1–5, 7–9, 11–12). Similar to ABR plants, inlet concentrations for COD and BOD exceeded 400 mg/L and 145 mg/L, respectively. outlet met COD and BOD guidelines in a limited number of plants and rounds (COD: P10: round 4; P11: rounds 3 and 5; BOD: P10: round 7; P12: round 9). Total nitrogen levels remained consistently high throughout all DEWATS plants. TSS levels complied with the guideline in specific plants and rounds (Plant P10: rounds 4, 7, 11; Plant P11: rounds 2–12; Plant P16: rounds 9–11). Phosphate levels adhered to regulations in some plants and rounds (Plant P10: rounds 7–12; Plant P11: rounds 3–12; Plant P12: rounds 3 and 9). Notably, pH levels generally remained within the acceptable range for both inlet and outlet samples.

The results for the LSP plants, depicted in Supplementary Figure S8 (Plants P4, P5, and P15) revealed varied patterns. For E. coli, counts were consistently high, except for the outlet in P4, which met the guideline with 0 cfu/100 mL in rounds 4 and 6. Helminth egg counts were within guidelines for P4 (round 10), P5 (round 12), and P15 (rounds 7 and 12). Outlet COD and BOD concentrations exceeded guidelines across all plants, except for the inlet BOD in P4 during round 5. Total nitrogen levels consistently remained above guidelines. TSS levels were within guidelines for P5 (rounds 8 and 11) and P15 (rounds 6 and 7). Phosphate levels were compliant in P4 (round 6), P5 (rounds 3–5), and P15 (rounds 1 and 5). pH levels were generally within guidelines, except for P4 (rounds 4 and 6).

UPF plants (plants P3, P6, P8, and P17, Supplementary Figure S7) also exhibited no clear trends in the data. Escherichia coli counts in the outlet met the guideline in several plants and rounds (Plant P3: rounds 1, 3, 9–12; Plant 6: round 4). Notably, helminth counts in the outlet were consistently below 150 eggs/L and adhered to the guidelines for specific plants and rounds (Plant P3: rounds 2, 6–8; Plant P6: round 6; Plant P17: rounds 5, 7, 9, 11). Outlet concentrations for COD and BOD met the guidelines in limited cases (COD: Plant P3: round 6; Plant P17: rounds 5 and 6; BOD: Plant P17: round 7). Total nitrogen levels remained persistently above the guideline across all UPF plants. TSS levels were within the limit for P3 (rounds 1 and 5–12), P6 (rounds 5 and 6), and P17 (rounds 5–12). Phosphate levels adhered to regulations in some plants and rounds (Plant P3: rounds 2, 4, 6; Plant P8: rounds 3–5; Plant P17: rounds 5–7; Plant P6: rounds 5 and 6). Generally, pH levels were compliant, except for some instances (Plant P1: round 2; Plant P7: round 3; Plant P6: round 12).

The results for the WSP plants, shown in Supplementary Figure S9 (plants P9, P13, and P14) exhibited similar inlet patterns. E. coli outlet counts met guidelines for P9 in rounds 4 and 7, and for P13 in rounds 4, 6, 7, 9, and 10. Helminth counts were within guidelines for P9 (rounds 2 and 8–12), P13 (rounds 1, 3, and 5–12), and P14 (rounds 7 and 12). Outlet COD and BOD concentrations were within guidelines for P9 in round 4, and for BOD only in round 5. P13 showed compliant levels for both COD and BOD in rounds 5–8. Total nitrogen levels were consistently above guidelines across all plants. TSS levels were below the guideline for P9 (rounds 4, 7, and 8) and P13 (rounds 1, 3, 5–8, and 10–12). Phosphate levels were compliant for P9 (rounds 7–9 and 11), P13 (rounds 3–8 and 10–12), and P14 (rounds 1, 5, and 12). pH levels were generally within guidelines, except for P13 in round 6.

The overall data distribution, including minimum, maximum, and quartiles, is presented in Table 4. For inlet samples, the highest median E. coli count was observed in WSP (4.15 × 10^6 cfu/100 mL), while the lowest was in ABR (9 × 10^5 cfu/100 mL). The interquartile range (IQR) revealed the greatest spread in UPF (1.18 × 10^7 cfu/100 mL) and the least in LSP (4.8 × 10^5 cfu/100 mL). In outlet samples, LSP exhibited the highest median value, whereas DEWATS showed the lowest (3,000 cfu/100 mL). The IQR was highest in UPF (2.6 × 10^5 cfu/100 mL) and lowest in WSP (22,375 cfu/100 mL). Violin plots (Figure 2) indicated that inlet E. coli concentrations were concentrated around 450,000 cfu/100 mL across all technologies, while outlet concentrations were around 20,000 cfu/100 mL for ABR, LSP, and UPF, and 3,000 cfu/100 mL for DEWATS and WSP.

For helminth eggs in inlet samples, DEWATS had the highest median value (19,325.5 eggs/L), while ABR had the lowest (6,250 eggs/L). The IQR was highest in DEWATS (1.1 × 10^5 eggs/L) and lowest in WSP (22,100 eggs/L). For outlet samples, WSP and DEWATS had the lowest median values (0 eggs/L), whereas LSP had the highest (50 eggs/L). Violin plots (Figure 2) showed inlet helminth egg concentrations were concentrated around 8,103 eggs/L, while outlet concentrations were below 20 eggs/L for ABR, DEWATS, and WSP, and around 55 eggs/L for LSP and UPF.

The highest median inlet COD concentration was found in UPF (4,910 mg/L) and the lowest in WSP (3,603 mg/L). The IQR was greatest in DEWATS (22,834 mg/L) and least in ABR (8,531 mg/L). In outlet samples, LSP had the highest median COD concentration (1,320 mg/L) and WSP the lowest (115 mg/L). Violin plots (Figure 2) showed inlet COD concentrations were concentrated around 2,981 mg/L, while outlet concentrations were below 1,097 mg/L for ABR, DEWATS, LSP, and UPF, and 148 mg/L for WSP.

The minimum and maximum median inlet BOD concentrations were observed in WSP (1,178 mg/L) and UPF (1745 mg/L), respectively. For outlet samples, WSP had the lowest median BOD concentration (36.5 mg/L), nearly within the Bangladesh standard, while LSP had the highest (366 mg/L). The IQR was highest in UPF and lowest in WSP. Violin plots (Figure 2) indicated inlet BOD concentrations were concentrated around 1808 mg/L, while outlet concentrations were above 244.7 mg/L for ABR, DEWATS, LSP, and UPF, and 33.1 mg/L for WSP.

For inlet TS concentrations, LSP exhibited the highest median (8,046.6 mg/L) and WSP the lowest (5,591.6 mg/L). The IQR was highest in DEWATS (20,949.7 mg/L) and lowest in UPF (7,363.6 mg/L). In outlet samples, DEWATS had the highest median TS concentration (4,087.4 mg/L) and WSP the lowest (870.9 mg/L). Violin plots (Figure 2) showed inlet TS concentrations were concentrated around 4,914.7 mg/L, while outlet concentrations were around 2,981 mg/L for ABR, DEWATS, LSP, and UPF, and 1,097 mg/L for WSP.

For inlet TSS samples, UPF had the highest median value (2,850 mg/L), and ABR the lowest (2,360 mg/L). The IQR was highest in DEWATS (19,975 mg/L) and lowest in ABR (6,793 mg/L). In outlet samples, LSP had the highest median value (415 mg/L) and WSP the lowest (43.5 mg/L). The IQR was highest in UPF (571 mg/L) and lowest in WSP (137 mg/L). Violin plots (Figure 2) indicated inlet TSS concentrations were concentrated around 2,981 mg/L, while outlet concentrations were around 148 mg/L for ABR, DEWATS, and UPF, and 1,097 mg/L for LSP, with less than 54.6 mg/L for WSP.

The highest median inlet TN concentration was observed in DEWATS (2,415.6 mg/L) and the lowest in WSP (1862.8 mg/L). The IQR was highest in DEWATS (1875 mg/L) and lowest in ABR (1,343 mg/L). For outlet samples, WSP had the lowest median TN concentration (115 mg/L) and LSP the highest (1,225 mg/L). The IQR was lowest in DEWATS (112.49 mg/L) and highest in UPF (201.79 mg/L). Violin plots (Figure 2) showed inlet TN concentrations were concentrated around 2,981 mg/L, while outlet concentrations were around 1808 mg/L for ABR, DEWATS, LSP, and UPF, and 121.5 mg/L for WSP.

For inlet PO₄ concentrations, DEWATS had the lowest median value (44.4 mg/L) and spread (63.4 mg/L), whereas UPF had the highest median (108 mg/L) and spread (91.9 mg/L). For outlet samples, WSP had the lowest median concentration (12.8 mg/L) and spread (16.7 mg/L), while LSP had the highest median (42.2 mg/L) and UPF had the highest spread (68.5 mg/L). Violin plots (Figure 2) showed inlet PO₄ concentrations were concentrated around 99.48 mg/L, while outlet concentrations were around 33.11 mg/L for ABR, DEWATS, LSP, and UPF, and 13.46 mg/L for WSP.

Inlet pH values ranged from 7.4 to 7.7 with minimal variation (IQR = 0.3–0.5). Outlet pH values were generally higher than inlet values for all technologies except DEWATS. LSP exhibited the widest pH range (7–12). Violin plots (Figure 2) indicated inlet pH concentrations around 7.7, while outlet pH concentrations were above 8 for all technologies except DEWATS.

COD removal efficiencies in various fecal sludge treatment technologies at Rohingya camps revealed that WSP were the most effective, achieving a 98.3% removal rate. DEWATS followed with a 94.4% removal rate, while UPF and ABR had similar efficiencies of 91.7% and 90.6%, respectively. LSP had the lowest efficiency at 81.6% (Figure 3). Seasonally, all technologies performed best during the monsoon, with WSP consistently showing the highest efficiency across all seasons and LSP the lowest, particularly in winter (Table 5).

For BOD removal, WSP had the highest efficiency at 97.8%, while LSP had the lowest at 80.9%. The removal efficiencies of DEWATS and UPF were similar, at 93.9% and 93.5%, respectively, with ABR achieving 88.5%. The mean BOD removal efficiency ranking was WSP > DEWATS > UPF > ABR > LSP (Figure 3). Seasonally, WSP maintained the highest BOD removal efficiency across all three seasons, with minimal variation. LSP consistently showed the lowest efficiency, particularly in winter (75%) compared to monsoon (91.3%) and summer (80.4%). There were no significant seasonal differences in the BOD removal efficiencies of DEWATS and ABR (Table 5).

All tested technologies effectively reduced TSS after the final treatment. WSP had the highest removal efficiency at 98.7%, while LSP was the least efficient at 86.3%. WSP, DEWATS, and ABR showed similar TSS removal efficiencies (Figure 3). Seasonally, WSP remained the most efficient across all seasons, with no significant variations. The other technologies performed best during the monsoon compared to winter and summer. LSP had a higher removal efficiency in monsoon (98%) and summer (97.4%), but significantly lower in winter (66.8%) (Table 5).

In terms of removal efficiency, WSP achieved the highest TS removal at 92.1%, while LSP had the lowest at 68.6% (Figure 3). The ranking of technologies based on TS removal efficiency is WSP > ABR > DEWATS > UPF > LSP (Figure 3). WSP consistently performed best across all seasons. Most technologies showed improved performance during the monsoon compared to winter and summer. UPF had the lowest removal efficiency in summer at 48.6%. LSP showed a significant drop in efficiency during winter at 51.7%. DEWATS and UPF performed better in the monsoon, while ABR had higher efficiency in summer than in the monsoon and winter (Table 5).

The WSP demonstrated the highest TN removal efficiency at 93.3%, outperforming other treatment technologies. Conversely, the UPF exhibited the lowest efficiency at 48.8%. Among the remaining technologies, no significant differences in removal efficiency were observed (Figure 3). Seasonally, WSP performed best during the monsoon, while DEWATS and LSP showed lower efficiency in the monsoon, summer, and winter, respectively. Notably, ABR maintained consistent TN removal efficiency across seasons, while DEWATS exhibited lower efficiency during the monsoon compared to summer and winter (Table 5).

Evaluating phosphate removal, WSP demonstrated the strongest capacity (82.8%), followed by UPF (63.7%), ABR (55.3%), DEWATS (53.7%), and lastly LSP at 49.2%. Notably, WSP and DEWATS maintained consistent performance across seasons, while UPF, ABR, and LSP exhibited variations, with UPF performing worst in summer and LSP in monsoon.

For the inlet samples, pH did not vary a lot. The median inlet pH value ranged from 7.4 to 7.7 with the least spread (IQR = 0.3–0.5) than other parameters. For the outlet samples, pH showed a comparatively higher value than their respective inlets for all the technology except DEWATS. On the other hand, LSP had the widest range which was from 7 to 12.

An evaluation of E. coli removal efficiency revealed UPF achieved the highest overall reduction (99.7%), followed by LSP (98.9%), ABR (98.6%), and WSP (97.3%). Notably, DEWATS displayed the lowest efficiency (75.6%). Seasonal variations impacted DEWATS and ABR significantly, with DEWATS dropping to 40.9% removal in monsoon compared to summer and winter. ABR also showed a substantial decrease in summer (82.3%) compared to other seasons. Conversely, WSP, UPF, and LSP maintained consistent performance throughout the year, with WSP achieving the best overall removal across all seasons.

WSP achieved the highest helminth removal efficiency at 100%, followed by DEWATS at 99.8% and ABR at 99.7%. UPF and LSP had lower efficiencies, at 93.7% and 82.3%, respectively (Figure 3). WSP consistently performed best across all seasons with 100% removal. DEWATS also maintained high efficiency (100% in summer and winter, 99.1% in monsoon), as did ABR (99.9% in summer, 99.8% in winter, 98% in monsoon). UPF was most effective in summer (99.8%), while LSP’s efficiency dropped significantly in winter (70.4%) compared to monsoon (95.9%) and summer (99.9%) (Table 5).

The outlet characteristics were compared with the Bangladesh Department of Environment (DoE) standards for sewage discharge to the receiving water bodies and land (The Environment Conservation Rules, 2023, 2023). For some parameters, Bangladesh does not have standard guidelines, in that case, outlets were compared with US, EU and WHO outlet/wastewater discharge guidelines (World Health Organization, 2006; Lu et al., 2016). Available guidelines are particularly not for the fecal sludge treatment plant outlets, but for a mixture of different wastewater including sewage discharge. Although it is evident that different technologies reduced the concentrations of most of the parameters from inlet to outlet, the concentrations recorded for BOD, COD, TN, TP, TSS, phosphate, E. coli and helminth eggs of the outlet exceeded the DoE discharge standards. However, some of the measured parameters during the study period were found to be within the guideline value (Figure 4).

In the case of COD and BOD, around 44.4% of the WSP outlet observations were within the guideline. However, none of the observations in LSP followed the standard guideline for BOD. In the case of pH, 100% of the observations of UPF and DEWATS were within the set guideline. In addition, pH of almost 95% of the observations in WSP and LSP followed the guideline value. In the case of phosphates, the highest percentages (61.1%) of the observations of WSP were within the standard guidelines followed by DEWATS (51.4%) and other technologies. In the case of TS, TSS and helminth, WSP showed better performances compared to other technologies based on standard guidelines (Figure 4). Almost 43% of outlet observations in DEWATS followed the discharge standard for E. coli. But for rest of the technologies it was less than 26%. For TN, none of the observations of four out of five technologies followed the standard discharge guideline and only 2.8% of the WSP followed the standard limit of outlet. Based on discharge guidelines, it can be said that WSP performed better than other FS treatment technologies in the camps. Technology-wise comparison of the percentages of outlet observations for a particular parameter within the standard guideline is depicted in Figure 4.

Correlation matrices for different FSM technologies present significant relationships among different FS quality parameters. Blank spaces denote non-significant correlations, blue-colored numbers indicate a significant positive monotonic correlation and orange-colored numbers indicate a significant negative monotonic correlation. The deeper the indicating color the stronger the correlation. The Correlation matrices are depicted in Figures 5A–E.

In ABR, E. coli was observed to have a moderate negative monotonic correlation with pH of the inlet whereas no correlation was found in the outlet. In the case of COD, the inlet had a very strong positive monotonic correlation with BOD, TSS, TS but in the outlet, the level of relationship demoted to a strong positive for both TSS and TS. Inlet BOD showed a very strong positive correlation with TSS and TS which decreased to a strong positive in the outlet (Figure 5A).

In DEWATS, the helminth of the inlet showed a strong positive correlation with COD, BOD, TSS, TS and TN whereas a weak positive correlation was observed with COD and PO₄ in the outlet. Though COD showed a very strong positive correlation with BOD, TSS, TS in the inlet that was reduced to a strong positive correlation in the outlet. The strength of the positive correlation of BOD with TSS and TS decreased from very strong to strong from inlet to outlet. TSS and TS also showed a similar trend in the inlet and outlet. TN showed a moderate negative correlation with pH and pH showed a weak negative correlation with PO4 in inlet whereas in outlet both correlations were positive (Figure 5B).

In the case of LSP, a strong positive correlation of helminth was found with BOD, COD, TSS, and TS in the inlet whereas in the outlet no such correlation was observed. Though inlet E. coli had a strong positive correlation with TSS and TS, in the outlet only PO₄ showed a moderate positive correlation with E. coli. Inlet COD and BOD had a very strong positive correlation with TSS and TS, whereas the level of correlation decreased in the outlet for each case (Figure 5C).

In UPF, E. coli showed a strong negative correlation with pH in the inlet whereas, the level of correlation decreased to moderate negative in the outlet. The level of correlation among COD, BOD and TSS decreases from inlet to outlet. Helminth showed a strong positive correlation with TSS in the outlet whereas no such correlation was observed in the inlet. Inlet pH showed a moderate negative correlation for COD, BOD, TSS and weak negative with TS whereas in outlet no such relation was observed (Figure 5D).

From WSP, the overall strength of correlation among different parameters decreases from inlet to outlet significantly. Inlet helminth showed a strong positive correlation with COD, BOD, TSS, TS and TN whereas no such correlation was observed in the outlet. Inlet COD, BOD, TSS and TS have a very strong to strong positive correlation among themselves whereas the level of correlation decreases in the outlet (Figure 5E).