Phnom Penh, the capital city of Cambodia (approximately 11°34″N, 104°55″E), is located on the Mekong floodplain, above the confluence of the Mekong, Tonle Sap and Bassac rivers (JICA, 2016). Phnom Penh has undergone rapid development and urbanisation in the past few decades. Recently, the whole city was divided into 14 districts, classified as urban areas (5 districts) and peri-urban areas (9 districts). The total land area of the city is about 679 km², with a population of approximately 2 million people in around 500,000 households (NIS, 2020).

Urban areas located in the centre of Phnom Penh are provided with full services in terms of water supply and sanitation (connection to sewerage network). The available network comprises a closed sewer system or an open canal system, depending on the location of the household within the city. Peri-urban areas can be described as adjoining areas, located outside formal urban boundaries and urban jurisdictions, that are in the process of urbanisation. These peri-urban areas can also be described as an interface, i.e., a transition zone or interactive zone, between urban and rural areas (Appiah-Effah et al., 2014b). Figure 1 shows a map of the study area, including the location of interviews, sampling sites and disposal sites for faecal sludge investigated in this study.

Phnom Penh still uses a combined drainage system that transports domestic, commercial and industrial wastewater, as well as stormwater flow during storm events. The combined wastewater is pumped into natural wetlands surrounding the city for treatment, before flowing to the final recipient waters (Mekong river and Tonle Sap river). There are two extensive wetlands that play key roles in treating wastewater from the whole city, Cheung Ek to the south of the city and Kob Srov to the north. However, the area of these wetlands is declining, due to the current rapid urbanisation and development in the city, as they are being filled with earth to reclaim land for development purposes (Doyle, 2013). Kob Srov wetland is a sewer entry point for Sen Sok district and high levels of untreated wastewater and faecal sludge are off-loaded into the wetland, accompanied by high levels of pathogens (Min, 2019). The Tonle Sap river is the final recipient of wastewater and faecal sludge from Kob Srov wetland.

Untreated faecal sludge can pose a significant health risk when dumped in the open environment, due to the presence of significant amount of bacteria, viruses and other pathogens (Strande et al., 2014). This is certainly the case in Phnom Penh, where downstream communities living along Tonle Sap are dependent on river water for their livelihoods and for key functions, including cooking and drinking and where river water contains varying levels of pathogens that carry risks of infection and illness (Min, 2019). Cheung Ek wetland, a seasonally inundated area located about 5 km to the south of Phnom Penh, receives around 80% of wastewater from Phnom Penh’s urban population and from factories (garment and others). This wetland is also used for aquatic plant and fish production, with harvesting being undertaken throughout the year.

Household survey: The household survey was designed to collect demographic information on on-site sewage containment users and to map the entire sanitation service chain, by tracking faecal sludge from source through emptying to the final disposal site. The survey was conducted in the period May-September 2020, and an attempt was made to include representative households in door-to-door data collection using a structured questionnaire. Households were selected based on information received from sewage emptying contractors about households requesting their services. These contractors normally offer two different types of service, either emptying sewage containers when full or de-clogging the containment/drainage network. Desludging is therefore included in both services. A total of 195 households were surveyed, representing both urban and peri-urban areas in Phnom Penh. Sampling was planned to collect proportional numbers of samples for urban and peri-urban areas, based on the local population in these areas. In total, 144 households in peri-urban areas and 51 households in urban areas were interviewed. Since the population in peri-urban versus urban areas in Phnom Penh is approximately 3:1 (NIS, 2020), the household sampling is representative.

The structured questionnaire included a combination of dichotomous, multiple choice and open-ended questions (see Supplementary Information). It was developed in English, before being translated into Khmer to simplify the interview sessions by using the local language. The questionnaire covered aspects of the household’s socioeconomic profile (including sex, education level, employment status, type of residential building, age of building, access to water), household sanitation practices (sewage container type and size, frequency of faecal sludge emptying, volume emptied) and householders’ perceptions of faecal sludge management. A draft questionnaire was pre-tested during 1 week at the beginning of the study and refined based on feedback from this field testing. A few modifications were made before the actual survey conducted. The final questionnaire version took around 20 min to complete and targeted any person in the household between 18 and 70 years old and aware of the sanitation system in the house. In most cases, the study team interviewed the head of the family. All households were allocated an identification code and the geo-coordinates (coordination system WGS 1984) of participating households were recorded using a handheld global positioning device (Garmin GPSMAP 60CSx).

Survey of emptying and transportation contractors: Another structured questionnaire was used for interviewing vacuum and manual sludge emptying contractors (see Supplementary Information). The purpose of interviewing contractors providing emptying services was to track the final fate of faecal sludge after removal from households. These interviewees were asked about the quantity of faecal sludge they collected and, where possible, the geo-coordinates (coordination system WGS 1984) of the disposal site of faecal sludge from each household was recorded using a handheld global positioning device (Garmin GPSMAP 60CSx). A specific name was assigned to each disposal site at which sludge was deposited. However, private contractors in Phnom Penh sometimes dump sludge illegally (Peal et al., 2015) and some sludge disposal sites had to be recorded as unknown, since a member of the study team was not allowed to accompany the truck driver to the disposal site in all cases.

Field observation: In addition to the interviews with householders and sewage emptying contractors, the study team observed the work performed by operatives during each emptying event. This allowed observations of the accessibility of the containers, respondents’ willingness to have faecal sludge treatment before final disposal, and whether the container emptying operatives used personal protection equipment while they performed the work. The study team also accompanied truck drivers to the disposal site and observed the surroundings at the sites, such as presence of water sources and the possibility of the neighbouring community using the site for swimming or for daily water extraction for general purposes.

Literature review: In addition to primary data collection, secondary data were collected from the literature in order to enable quantification of faecal sludge and resources. Data sources included government reports on population census, published literature on the population served by on-site sanitation in Phnom Penh and published information on average urine and faeces generation rates in the city. Statistical data from the Food and Agriculture Organization (FAO) on the total nutrient content in staple foods consumed by Cambodians were used to calculate the nutrient content in combined excreta and in faecal sludge.

Statistical analysis: Microsoft Excel 2010 and R software version 4.0.4 were used for data handling and analysis. Descriptive statistics were calculated, such as proportion test and Chi-square test/Fisher’s exact test (where the number of samples (n) broke the rule of thumb that n (1-p) >10. Samples must be taken for household data to reveal socio-economic status in relation to sanitation practices at household level, especially as regards FSM. p-values <0.05 were considered statistically significant.

Spatial analysis of faecal sludge disposal sites: Geo-coordinate data on households and sludge disposal sites were processed using Microsoft Excel 2010. The distance from each household to its sludge disposal site was calculated using ArcMap 10.8. The drainage network system serving households within the coverage area was used to identify the final disposal site (recipient waters) for faecal sludge. The linear distance calculation method was used to estimate the distance between source household and final sludge disposal site. Three transport zones (4, 9 and 14 km) were added to the map to assess the distance between the two main disposal sites and the households from which the faecal sludge was obtained.

Faecal sludge quantification: The sludge collection method developed by (Strande et al., 2014) was used to quantify the amount of faecal sludge handled throughout the entire sanitation service chain. Based on population data for 2020, the amounts were quantified at six different stages of the chain, using a modified approach taken from Strande et al. (2018). The parameters determined at these stages were excreta generation rate (Q₁), faecal sludge generation rate (Q₂), faecal sludge accumulation rate (Q₃), amount of faecal sludge emptied (Q₄), amount of faecal sludge collected and delivered to Boeung Trabek pumping station (Cheung Ek wetland) (Q₅), and amount of faecal sludge collected and delivered to Prek Pnov open canal (Kob Srov wetland) (Q₆).

Q₁ was calculated as:

Excreta produced Q 1 ( L/year ) =  P ( served ) x ( Q ( urine )   +  Q ( faeces ) ) (1) where P_(served) is the population served by on-site sanitation in Phnom Penh; Q_(urine) is urine generation rate, which was set at 1.42 L/cap/day (Rose et al., 2015); and Q_(faeces) is estimated faecal generation rate, set at 0.236 L/cap/day for low-income countries (Strande et al., 2018).

Q₂ was calculated as:

Faecal sludge produced Q 2 ( L/year ) =  Q 1   +  Total container inflow (septic tank +pit latrine) (2) where: total container inflow ( septic tank + pit latrine ) = P ( served )   x   C w   (3) and C_w is the quantity of water inflow to the container (septic tank and cesspit). Key assumptions made were 1) that water inflow is similar for septic tanks and latrines, 2) that type of container does not influence faecal sludge characteristics (based on Eliyan et al. (2022)); and 3) that water and excreta are the only substances entering the container, since water is used for anal cleansing and households predominantly have a piped water connection, while the small proportion of the population that use toilet tissue for wiping usually dispose of it in trash bins with other types of solid waste. According to Koppelaar et al. (2018), an average of 58.6 L/cap/day of water enter the sewage container (C_w) in developing countries.

Q₃ was calculated as:

Faecal sludge accumulation Q 3 ( L/cap/year ) = E m p t i e d   v o l u m e N u m b e r   o f   u s e r s . E m p t y i n g   f r e q u e n c y (4)

The input values used for calculating Q₃, i.e., emptied volume, number of users and emptying frequency, were the average value for each category based on the household questionnaire and triangulated with data from the container emptying contractors.

The amount of faecal sludge emptied (Q₄) was calculated based on observations during each emptying event. All faecal sludge in the container was removed and only a small amount of water was sprayed to clean the container, so it was assumed that faecal sludge emptied (Q₄) was equal to faecal accumulation rate (Q₃). The analysis covered only faecal sludge collected by mechanical emptying contractors.

Faecal sludge collected and delivered to Cheung Ek wetland (Q₅) was estimated as the amount of sludge collected from household containers and delivered to the authorised disposal site. According to Peal et al. (2015), Boeung Trabek pumping station is the only authorised disposal site for Phnom Penh. Therefore Q₅ was determined based on data collected from the interviews with container emptying contractors on whether they discharge the sludge they collect at Boeung Trabek pumping station or directly into Cheung Ek wetland. Q₆ was defined similarly as the amount of faecal sludge collected from households and discharged into Toul Sampov wastewater canal or Kob Srov wetland, based on response from contractors during interviews and on field observations. Toul Sampov canal, which is located to the north of the city (see Figure 1), is 5 km long and carries wastewater from the Sen Sok area to Kob Srov wetland.

Resources quantification: Resources can be described as the amount of nutrients and energy that could be recovered from faecal sludge. According to FAO (2019), the total protein content in food consumed by the Cambodian population is 65.53 g/cap/day and the protein content in vegetable products consumed is 46.81 g/cap/day. The total amounts of the macronutrients nitrogen (N) and phosphorus (P) in faecal sludge in Phnom Penh were calculated using Eqs. 5 and 6, respectively (Jönsson et al., 2004) and considering the fact that only 22% of on-site sanitation users report employing a contractor to empty their sewage container (Frenoux et al., 2011). Content of nitrogen ( N ) =  0 . 13   x   Total food protein (5) Content of phosphorus ( P ) =  0 . 011 x ( Total food protein + vegetable food protein ) (6)

The nutrient resource in faecal sludge was also calculated based on concentration of total nitrogen (N_tot) and total phosphorus (P_tot) in faecal sludge according to (Eliyan et al., 2022).

The potential for energy generation from faecal sludge was estimated based on Ahmed et al. (2019), who concluded that the energy potential in faecal sludge lies within the range 16.39–18.31 MJ/kg at a sludge density of 1,001 kg/m³ (Radford and Sugden, 2014).

There was no statistical correlation between demographics of the respondents and geographical locations (see Supplementary Table S1).

Two types of on-site sewage containment system are used in Phnom Penh, cesspits and septic tanks, According to our survey of households, cesspits dominate, serving up to 92.8% of the population, a trend seen in both urban and peri-urban areas. Around 95% of urban households reported having their sludge container connected to the sewer network, while only 62.5% of households in peri-urban areas reported have a direct connection (p < 0.001). Concerning the sanitation management practices performed by respondents, type of containment system and type of wastewater received by the system (only blackwater or not) were found to be unaffected by location in urban or peri-urban areas in Phnom Penh (Table 1). However, the age of the sewage container differed significantly with the geographical location of the household. The containers at houses in peri-urban areas of the city tended to be newer, reflecting the fact that the city is developing and expanding outwards. Mechanical emptying services is the only preferred method for households in Phnom Penh when their containments were full or clogged. No evidence of manual emptying practices was found. According to the observation by the study team during data collection, none of pit emptiers used personal protective equipment during emptying events. Hence, it might potentially pose risks to their health.

There is no faecal sludge treatment facility in Phnom Penh and Boeung Trabek pumping station is the only authorised sewage disposal site (Peal et al., 2015; JICA, 2016). The disposal sites identified in this study included public manholes near the households where faecal sludge was collected, fields around the Kob Srov area, Toul Sampov wastewater canal, a smaller canal (1 km) connected to Toul Sampov wastewater canal, and Boeung Trabek pumping station (Cheung Ek wetland). The survey also revealed that Cheung Ek wetland is the main disposal site (receiving 54.1% of all sludge collected), followed by the small canal and Toul Sampov wastewater canal itself (34.5%). Toul Sampov canal receives wastewater from the Sen Sok area, which flows onwards by gravity to Kob Srov wetland, with the Tonle Sap river being the final receiving reservoir. The remaining 11.4% of collected faecal sludge goes to open fields in the Kob Srov area and public manholes near source households. Since those two main disposal sites are pumping stations, there is limited risk for spillage and spread of faecal matter to local people living around those areas.

Wherever faecal sludge is disposed of within the drainage network, it ends up in one of the two main receiving wetlands, namely Cheung Ek and Kob Srov. The results obtained in this study indicated that Cheung Ek wetland is the main faecal sludge disposal site for container-emptying contractors (57.8%), while Kob Srov receives 42.2% of all sludge collected from household sewage containers by mechanical emptiers (Table 2). The mean travel distance from source households to Cheung Ek was found to be 4.34 km, while that from source households to Kob Srov was around 3.87 km. The shortest estimated distance observed was 0 km, in cases where the faecal sludge removed from a household’s containment system was disposed of in a manhole located in front of the household. This only occurred for households with drainage network coverage.

The linear distance from source (extraction household) to each disposal site was used to estimate the travel distance for discharging emptied faecal sludge from households in Phnom Penh. Three zones were created around the two main disposal sites, to group travel distances for emptying events. The resulting map revealed that most travel distances for emptying faecal sludge fell within the first and second zones, with few distances within the third outer zone (Figure 2). This reflects the current practice of contractors, who prefer not to travel long distances to discharge collected sludge when there is an opportunity to dispose of it somewhere that could reduce their travel distance, thereby saving transportation time and fuel costs.

Estimation of excreta production (Q₁) and faecal sludge generation (Q₂) was based on secondary data taken from the literature, based on Strande et al. (2014) as indicated in data analysis section. The production rate of excreta in Phnom Penh was taken to be 604 L/cap/year for all types of containment system, based on findings (Eliyan et al., 2022) that type of containment system does not influence the characteristics of faecal sludge. The faecal sludge generation rate (Q₂) was estimated to be 21,993 L/cap/year. Based on the primary data collected in the study, faecal sludge accumulation (Q₃) was estimated to occur at a rate of 106 L/cap/year for all types of containment system. This was only around half the value reported previously for the city of Kampala in Uganda (Strande et al., 2018). However, an earlier study conducted in 12 Asia and Africa cities found faecal sludge accumulation rates varying from 35.6 to 959 L/cap/year (Chowdhry and Kone, 2012). The accumulation rates in Phnom Penh are at the lower end of that reported range, possibly because containment systems in Phnom Penh are usually connected to the sewerage network, which allows daily overflow of supernatant from the sludge container to the drain network. In addition, many of the household containment systems in the city are not watertight, which allows the liquid portion of wastewater in the container to drain out to surrounding soil. According to our calculations, the total amount of faecal sludge emptied (Q₄), and thus collected (Q₅), was 32,500 m³/year (Table 3).

Our calculations showed that around 52.5% (18,800 m³/year) of total faecal sludge emptied from household containment systems during the study period was taken to Cheung Ek wetland and 42.2% (13,700 m³/year) was discharged in the Kob Srov catchment.

Potential resources assessed in this study were the amount of nitrogen, phosphorus and energy contained in faecal sludge. Based on FAO protein consumption data and resulting N and P in excreta (Eqs. 5 and 6), it is estimated that each individual excretes around 3.12 kg N and 0.45 kg P per year (Table 4), the total amount of nitrogen theoretically present in excreta (urine plus faeces) was thus estimated to be 955 tons/year, while the amount that could potentially be extracted from faecal sludge was only 6 tons/year (Jönsson et al., 2004). Thus, according to these findings faecal sludge in Phnom Penh contains less than 1% of total nitrogen excreted by humans. Nitrogen in wastewater is mostly found in the water-soluble form as ammonia and follow the liquid fraction into the sewer network or into the ground due to non-watertight containers.

The results for phosphorus showed that a larger fraction, around 9%, remains in faecal sludge (Table 4), presumably because phosphorus tends to precipitate as metal phosphate and attach to solid particles in sludge and is less water-soluble than nitrogen. However, a high proportion of both nutrients (nitrogen and phosphorus) remains in the liquid wastewater fraction, which with improved wastewater treatment could be captured and treated as part of achieving the UN SDG goal 6, under target 6.2 and 6.3, as well as meeting the Cambodian wastewater discharge standard (RGC 2017; RGC, 2021) and to avoid environmental impacts. In conclusion, around 6 tons of nitrogen and 13 tons of phosphorus could be recovered from faecal sludge annually.

Potential energy generation was calculated based on the total faecal emptied annually (Q₄). Based on energy potential from Ahmed et al. (2019), the estimated amount of potential energy that could be captured from faecal sludge annually was within the range 532,571–594,959GJ, or 148-165 GWh.