Off-grid rainwater and greywater treatment and reuse on household level: Conceptual approach and pilot operation at the KREIS-Haus demonstration case, Switzerland

The fraction of the global urban population facing water scarcity is projected to increase from one-third in 2016 to nearly half in 2050 (He et al. 2021). Even in countries such as Switzerland, traditionally not associated with water scarcity, climate change is causing regional and seasonal water shortages by altering rainfall patterns (Fischer et al. 2015) and reducing glacier and snow cover (Brunner et al. 2019). The implementation of local water recycling is a possible approach to managing water scarcity while increasing resilience (Rozos et al. 2010). If correctly implemented, water reuse could foster energy and water conservation, recovery of by-products, avoid discharge of pollutants to the environment, yield economic benefits, and have beneficial impacts on local communities and the environment (Cagno et al. 2022). The principle of separating different waste streams and water qualities for appropriate treatment and reuse (Otterpohl et al. 2004) is applicable at various scales, from urban infrastructure integration (Crosson 2021) to neighbourhoods (García-Montoya et al. 2016) and at the building level (Muthukumaran et al. 2011), ultimately leading to fully off-grid, self-sufficient systems. Hoffmann et al. (2020) categorize wastewater systems into grid-dominated, non-grid, small-grid, and hybrid solutions, where non-grid systems avoid the need for sewer connections. Self-sufficiency is the ability of a system to meet its own needs independently, without relying on external resources. In this context, water self-sufficiency refers to the ability to fulfil water demands using locally sourced water, such as rainwater harvesting, water reuse, or desalinated water (Rygaard et al. 2011). Off-grid water management is increasingly recognized as a solution to the challenges of centralized systems, such as aging infrastructure, high maintenance costs, and limited scalability, by enhancing resilience through decentralized water supply and reduced reliance on centralized infrastructure (Hoffmann et al. 2020).

Decentralized systems apply advanced technologies to achieve water and energy self-sufficiency. Key innovations include rainwater harvesting, greywater recycling, and advanced treatment methods like membrane bioreactors and electrolysis, enabling on-site water reuse for domestic purposes (Maniam et al. 2022). Decentralized water reclamation systems can integrate resource recovery with water reuse, achieving considerable cost savings and energy self-sufficiency at the neighbourhood scale through technologies like anaerobic membrane bioreactors and biogas digesters (Al-Azzawi et al. 2022). Photovoltaic energy generation and vertical farming further enhance self-sufficiency and resource optimization (Garrido-Baserba et al. 2024).

In their review, Maniam et al. (2022) discuss important considerations for decentralized water reuse systems. These include selecting appropriate technologies tailored to specific applications, ensuring cost feasibility, adhering to water quality standards, and addressing public acceptance through stakeholder engagement. They also discuss the importance of energy-efficient practices, regulatory support, and adapting technologies to regional contexts, which are essential for the long-term viability of such systems. To achieve energy and cost efficiency while maintaining adherence to water quality standards, the concept of ‘fit-for-purpose’ is of particular interest (Capodaglio 2021). This approach ensures that water is treated to meet specific quality standards for each reuse objective. By aligning treatment levels with intended uses, this strategy can save water while also reducing production costs and energy demand (Capodaglio 2021).

While many case studies explore aspects of decentralized water treatment systems, such as greywater reuse or rainwater harvesting, case studies in Europe on fully water-self-sufficient, off-grid systems that combine such approaches, remain limited. A deeper understanding is needed of the long-term feasibility and operational performance of these systems, particularly under varying user behaviours, climatic conditions, and occupancy patterns.

The KREIS-Haus in Feldbach, Switzerland, fills this gap. It is a demonstration tiny house established in 2021, functioning as a ‘real-life laboratory’ to explore sustainable building practises (Buehler et al. 2025). The name KREIS-Haus is derived from the German abbreviation for ‘Energie- und Ressourceneffizientes Suffizienz-Haus’ (‘Energy and Resource Efficient Sufficiency House’). It consists of a living unit with a kitchen and bathroom imbedded into a greenhouse, which plays a crucial role in energy provision and closing the water and nutrient cycle, and also serves as a garden and living area. The implemented water cycle integrates low-tech water treatment technologies for the (re-)use of rain and greywater within a residential setting. Low-tech water treatment systems prioritize affordability, simplicity, durability, and low and simple maintenance (McBean 2024). KREIS-Haus is distinctive for integrating a comprehensive range of sustainable building practices within a small footprint, including circular water management with renewable energy provision, sustainable building materials, circular building techniques, and space-saving design. KREIS-Haus is open to the public for overnight stays, allowing a wide range of users to experience living in a sustainable house and providing insights into the system's performance under different user behaviours.

In Switzerland, the use of alternative water sources, such as rain and greywater at the household level, is not widely practiced. While rainwater is occasionally utilized, such as in alpine huts, the reuse of treated greywater remains rare. Therefore, Knabl et al. (2024) emphasize the need for more pilot projects to gather local experiences and develop guidelines tailored to the Swiss context. Furthermore, Craig & Richman (2018) stress the importance of a comprehensive evaluation, encompassing not only water balance and water quality but also aspects such as energy consumption, durability, maintenance requirements, installation process, economics, and user satisfaction. This study addresses this gap by assessing the technical and economic viability of the KREIS-Haus water system through an analysis of water mass balance, water quality parameters, compliance with water quality guidelines, energy consumption, and cost efficiency.

The electricity for the KREIS-Haus, including the water treatment, is generated by two types of solar modules integrated into the saddle roof: 4 kWp semi-transparent glass modules (Ertex Solartechnik GmbH, Amstetten, Austria) on the south side and 6 kWp Sunskin roof 145/195 glass modules (Swisspearl Schweiz AG, Niederurnen, Switzerland) on the north side. Excess electricity is stored in LiFePO4 second-life batteries (Kyburz Switzerland AG, Freienstein, Switzerland) with a capacity of 14.4 kWh or fed into the grid when the batteries are full. If the solar system and batteries do not supply enough electricity, additional power is drawn from the grid.

The rainwater is collected from the saddle roof with a 78 m² base area, passes a basket sieve that retains coarse impurities and is then stored in a 3,000 L rainwater tank (Figure 1). When the rainwater tank is full, the water overflows and infiltrates into the ground. When potable water is needed in the house, a submersed pressure pump pumps the water through a sediment filter (Pentair DGD-2501 25/1 μm, Lenntech B.V., Delfgauw, The Netherlands), a granular activated carbon filter (Pentair GAC-20BB, also Lenntech B.V.), and an ultraviolet light-emitting diode (UV-LED) lamp (PearlAqua Deca, AquiSense, Erlanger, USA), which treats up to 45 L/min. The treated rainwater is used for the shower, kitchen, and bathroom sink, which are all equipped with highly water-efficient faucets. The untreated rainwater is used for irrigation of half of the rooftop garden.

The wastewater from the kitchen and bathroom flows into the greywater tank of 1,500 L. From there, it is pumped once a day into the greywater treatment unit (own design, produced by Abderhalden Gartenbau AG, Wildhaus, Switzerland). The treatment system is housed in a box with external dimensions of 2.22 × 0.82 × 1.53 m (L × W × H) and holds a total water volume of 1,484 L (Supplementary Material, Figure S1). Its design is based on prior research on bio pools (Frei & Antenen 2021) and treatment of laundry effluent (Buehler et al. 2021), with a submerged fixed-bed biofilm reactor (Schlegel & Koeser 2007) as core technology. The treatment unit encompasses four sequential chambers: a settling tank (497 L), two biofilter chambers (each 247 L) filled with open-structured plastic media and equipped with an aeration pump (Eco Air Pump AP35 20 W, Aqua Forte, SIBO Fluidra, Veghel, Netherlands) with a capacity of 35 L/min, and a treated water storage chamber (501 L) with a submerged UV lamp (Light Tech T5 Amalgan 40 W, teich-direct, Rheine-Elte, Germany). The greywater treatment unit is fed by two pumps. The first pump transfers the greywater from the greywater tank to the first chamber, with a flow rate of 8.56 L/min for 23 min/day. The second pump transfers the water from the third to the fourth chamber with a flow rate of 16.56 L/min for 12 min/day. The water flows passively between the first, second, and second and third chambers. Based on these flow rates and operational timings, the system treated 200 L/day. The treated greywater was used in the house for the washing machine (Spirit 520, Schulthess Maschinen AG, Cham, Switzerland) and irrigation of the other half of the rooftop garden. Unused treated greywater flows back into the greywater tank. When the greywater tank is full, the water flows into the overflow tank (1,500 L), which needs to be manually emptied. All three tanks are buried in the ground under the house. They are made of plastic and of the type of Graf Platin flat tank (Otto Graf GmbH, Teningen, Germany). The water in the greywater tank is, therefore, a combination of untreated and treated greywater. However, for simplicity, the water in the greywater tank will be referred to as ‘greywater’ throughout this study.

Prior to this study, the house was intermittently inhabited, and the treatment system was operational for nine months. During this study, weekly water samples were collected at four sampling points, i.e. before and after treatment for both rainwater and greywater, from 22 July 2022 to 24 April 2023. Samples of the overflow tank were taken four times over the entire period. The greywater sample was collected directly from the greywater tank, while all other samples were taken from the respective taps within the house (Figure 1). On four occasions, additional samples were collected at sampling points 1 and 2 for an in-depth analysis of the rainwater treatment, comparable to a drinking water assessment.

On-site measurements of turbidity were conducted using a portable turbidimeter (Hach 2100Qis, Hach, Loveland, CO, USA), while pH, electrical conductivity (EC), dissolved oxygen (O₂), and temperature were determined using a portable multimeter (Hach HQ40D). Samples for further analysis of abiotic parameters were collected in plastic tubes and frozen prior to analysis, whereas samples for biological parameters were collected in glass bottles and analysed on the same day.

Analysis of chemical oxygen demand (COD), total phosphorus (TP), total nitrogen (TN), and cationic and anionic surfactants was done spectrophotometrically (Hach LCI 400 for COD, Hach LCK 348/LCK 349 for TP, Hach LCK 138 for TN, Hach LCK 331 and LCK 332 for the respective surfactants). Biological oxygen demand (BOD5) was measured using the standard method APHA 5210-B only at sampling points 3 and 4. Additionally, COD, TP, and TN were measured in the discharge of the overflow tank at sampling point 5.

Microbiological parameters (enterococci, total aerobic count, Escherichia coli, and total coliforms) were analysed using compact dry plates (Compact Dry ETC 1002945, TC 1000167, and EC 1000169, Shimadzu Diagnostics Corporation, Tokyo, Japan) at sampling points 1–4. For Pseudomonas aeruginosa, CHROMID 43,462 plates (bioMérieux SA, Marcy-l'Étoile, France) were used. The preparation of E. coli, coliforms, and enterococci followed the membrane filter method (HyServe no date), in which 100 mL of the sample was filtered (membrane size 0.22 μm), and the filter was then placed onto the corresponding plate. For total aerobic count 1 mL, and for P. aeruginosa 0.1 mL was directly used as inoculum onto the plates. Dilution series were performed as needed. E. coli, coliforms, and enterococci were cultivated at 35 °C for 24 h according to the manufacturer's specifications. Total aerobic count and P. aeruginosa were cultivated at 30 °C for 72 h, following the instructions of the Swiss drinking water regulation (EDI 2016). Legionella was analysed according to DIN EN ISO 11731:2019-03 at sampling points 3 and 4.

The occupancy of the house was calculated based on the number of occupied nights in the preceding week, irrespective of the number of guests staying. In the shower, eco-friendly body care products were provided, though their use was not mandatory. For dishwashing, handwashing, and laundry, a pure-castile multi-purpose liquid soap was used. Laundry was done without the use of softener. All cleaning products were eco-friendly.

Water consumption data were collected using two water flow meters (FCH-C-Ms-N, Bio-Tec, Vilshofen, Germany) from 1 July 2023 to 29 February 2024 in 15-min intervals (Figure 1). The energy consumption of the water system was logged with an MDT AMI-0816.02 actuator from 10 July 2023 to 29 February 2024.

To assess the water mass balance, 12 water flows within the house, labelled A to L, were monitored or calculated (Figure 1, Supplementary Material, Table S1). Flows C, E, F, and K were directly logged by water meters, with flows C and E and F and K being measured by the same meters. The specific water usage for flows K and E (irrigation) and F (laundry) was determined based on whether the power was on for those operations. The remaining flows were calculated as detailed in Table S1. Additionally, a simulation using Stella Architect 3.2.1 (isee systems, inc., Lebanon, New Hampshire, USA) modelled continuous occupancy of two persons over a 1-year period. For this, occupancy rates were adjusted accordingly, and monthly rainfall data for one year (March 2023 to February 2024) from MeteoSchweiz (2024) were applied. Average daily water consumption was calculated based on the data collected during the monitored period. The water mass balance for both the sampling period results and the simulation was visualized using the software e!Sankey 5.2.1 (iPoint-systems GmbH, Hamburg, Germany). The level of water self-sufficiency was calculated using the method proposed by Rygaard et al. (2011), defined as Qlr/Qtd, where Qlr represents the amount of locally sourced water, rainwater, and treated greywater in this case, and Qtd is the building's total water demand.

Prior to data analysis, a data cleaning step was executed. For abiotic parameters, results below the limit of detection (LOD, as specified by the individual analytical methods) were substituted with half of the LOD. For biotic parameters, measurements obtained from dilutions of 1/10 and higher that resulted in counts below the threshold of 10 counts per plate were excluded from the dataset. Outliers were identified and removed from the data set based on the three-sigma rule (Li & Chen 2012). Outliers reflecting real system variations were retained, while those likely caused by measurement errors (e.g. implausible or physically impossible values) were removed. Decisions were based on contextual analysis and checks for physical plausibility.

Removal rates were calculated by using the mean values of the whole sampling period for the parameters before and after treatment. For a further analysis of the greywater treatment performance, the dataset was divided into two subsets: ‘Low Occupancy’ (1–4 nights/week) and ‘High Occupancy’ (5–7 nights/week), using the occupancy as specified in section 2.2. This resulted in an equal number of data points in each subset.

The analysis of the energy balance focused on the electrical energy consumed by the water treatment system, including the booster pumps required to supply treated water to the house. Other energy uses, such as those for heating water, which are also present in conventional households, were not included in the scope of this study. The energy consumption was analysed using two methods. The first method was based on measured values of energy consumption, encompassing both the rain and greywater treatment systems, from 10 July 2023 to 29 February 2024. The second method calculated the energy consumption for the rainwater and greywater treatment systems separately. This involved determining the power rating (in Watts) from the energy meter for each component and multiplying it by its estimated daily runtime (in hours). The results of both were compared to similar systems from the literature. This dual approach enabled the evaluation of actual metered values for the entire system while also allowing for comparison and benchmarking of the separate rainwater and greywater treatment systems. For the greywater treatment system, a calculation incorporating a series of optimization measures was conducted, including treatment capacity expansion to 480 L/day, pump reduction, a simplified control system, and replacing the UV lamp with a more efficient UV-LED lamp (Supplementary Material, Table S6).

To assess the economic feasibility of the KREIS-Haus water concept over traditional centralized water infrastructure, a cost analysis was performed. Table 1 outlines the six scenarios evaluated (S1–S6), two for the KREIS-Haus (S1, S2), and four for a centralized approach (S3–S6), each considering a household of four persons over a 30-year lifespan. Assumptions included static municipal costs for freshwater supply, wastewater treatment, and electricity. Although KREIS-Haus is in its current form designed for an occupancy of two persons, a four-person household was selected to align with a more sustainable scenario, where a single-family home accommodates more occupants. Since the cost analysis focuses solely on the water system and not the entire house, the investment costs would remain largely unchanged if implemented for a four-person household with the same sanitary installations (one bathroom, one kitchen). Minor cost increases due to slightly longer pipes or larger tanks are already accounted for in the assumed investment cost of CHF 25,000, which reflects an estimated future market price for the system rather than a detailed cost calculation.

The two KREIS-Haus scenarios entailed one where no connection fee would be required (reflective of KREIS-Haus's self-contained system) (S2) and another scenario where a connection fee was paid (S1) in accordance with current Swiss regulations, which generally do not allow for fee exemptions and the absence of a physical connection. Costs for KREIS-Haus also included the grid electricity price, despite most electricity being sourced from the in-house solar plant. The system's reduced electricity consumption of 1.25 kWh/day was used for greywater treatment. Additional costs included regular maintenance such as cleaning the rainwater tank, replacing the rainwater treatment filters, and backflushing the biofilter of the greywater treatment, assumed to require an average of 10 h annually at a rate of 80 CHF/h based on the experience of operating the KREIS-Haus. Annual component replacement costs were estimated at 2% of the initial purchase price, which encompasses the full water system setup within KREIS-Haus, projected at a future market rate. Details on the costs for maintenance materials can be found in the Supplementary Material, Table S3.

The four scenarios based on a centralized approach incorporated user costs for connection fees and ongoing freshwater and wastewater charges. Scenario S3 reflects the current conditions at the location of KREIS-Haus in the municipality of Hombrechtikon ZH. Scenario S4 assumed increased water fees akin to the highest rates within Swiss municipalities. Scenario S5 assumed the absence of infrastructure, necessitating a 50-m pipeline installation at the homeowner's expense. Scenario S6 combined the assumptions of scenarios S4 and S5.

A sensitivity analysis was conducted considering these input parameters: an electricity price of 0.3 CHF/kWh, a household with two occupants, 20 h of maintenance per year, an hourly maintenance rate of CHF 40, a combined scenario of 20 h of maintenance at a CHF 40 hourly rate, and a lifespan of 20 and 40 years. The detailed calculations and the results of the sensitivity analysis are in the Supplementary Material, Tables S3 and S4.

The simulation for one year under full occupancy of two persons projected an increase in rainwater utilization to 62%, meaning that under this scenario complete water self-sufficiency was maintained. In contrast to the Swiss average water consumption of 142 L per person per day (SVGW 2018), residents of the KREIS-Haus used an average of 66 L treated rainwater per day. This is a result of employing a dry toilet, recycling treated greywater, and using water-efficient faucets. In addition, the low water consumption may also be influenced by the house's use as a holiday home, where user behaviour, such as less cooking and spending less time in the house, differs from that of a typical residential home.

Furthermore, the simulation projected that the house could operate fully self-sufficiently with up to three occupants. However, the simulation was based on monthly rainfall data from 2023/2024. Given the effects of climate change, which are in Switzerland altering rainfall patterns (Fischer et al. 2015), longer dry periods could compromise the system's ability to maintain complete self-sufficiency. To address this, the system could implement a larger rainwater storage tank, expand the catchment area by incorporating nearby roofs, such as carports, and increase the reuse of treated greywater for applications closer to potable standards, such as bathroom use, with the addition of an appropriate treatment step.

The high self-sufficiency of the KREIS-Haus exceeds values from other case studies, such as 15% in Orange County, California, and up to 80% in Pimpama-Coomera, Australia (Rygaard et al. 2011). This demonstrates the potential of decentralized systems to achieve full water self-sufficiency.

The rainwater quality was compared before and after treatment (Table 2), along with a comprehensive analysis of additional parameters (Supplementary Material, Table S2) to evaluate its suitability for human consumption. In terms of pollution, the rainwater fell within the middle range compared to other studies; for example, COD was 13.2 mg/L, falling within the range of 0.1–49.25 mg/L reported by Liu et al. (2021). The turbidity was measured at 2.7 FNU, while other studies reported values between 0.3 and 5 NTU (García-Ávila et al. 2023), though the non-comparable units (FNU/NTU) should be noted. The pH of 6.4 was on the lower end of the range reported in other studies, which varied between 6.3 and 8.4 (Shaheed et al. 2017; García-Ávila et al. 2023), suggesting the rainwater was more acidic than in other locations but still above the threshold of 5.6 to be classified as acid rain (Keresztesi et al. 2019). Furthermore, chrome, nickel, and copper were present, albeit in very low concentrations. Both the acidity and the presence of metals may be due to internal or external factors. External sources include the deposition of air pollutants generated by traffic and fuel combustion or from industrial activities (Keresztesi et al. 2019; Wartalska et al. 2024). Internally, the metals could be attributed to leaching from the copper gutter, piping, and other metal-containing parts of the water system of the KREIS-Haus (McIntyre et al. 2019), likely influenced by the lower pH, which can accelerate metal leaching (Wang et al. 2014). The acidity could result from storing the rainwater in a plastic tank, which may contribute to more acidic conditions compared to a concrete tank (Wartalska et al. 2024).

E. coli was not detected in either the raw or treated rainwater, consistent with findings in five out of eight studies reviewed by García-Ávila et al. (2023). Removal rates of 29% for COD and 27% for turbidity were achieved, along with reductions ranging from log 0.3 for coliforms to log 1.3 for total aerobic count (Table 2). In comparison, Gabriela & Vladimir (2022) found that similar treatment methods, such as granular filtration and microfiltration, achieved turbidity removal rates between 13 and 99%. Additionally, the removal of enterococci was relatively low and insufficient, potentially due to their higher resistance to the UV LED light used in the treatment. Similarly, several studies have reported that enterococci are more resistant than other pathogens (Fiore et al. 2019; Fiorentino et al. 2021). To further evaluate the potential influence of water stagnation, we assessed the risk of pathogen regrowth during prolonged periods of non-use. Given the short length of the pipe (approximately 2 m) transporting treated rainwater from the UV disinfection unit to the kitchen and bathroom taps, which represents an almost point-of-use disinfection setup, the risk of stagnation and associated regrowth is minimal.

The greywater treatment unit's performance was evaluated by comparing water quality parameters before and after treatment (Table 2). High removal efficiencies were observed for COD (88%), BOD₅ (85%), and turbidity (97%), while nutrient removal was lower, with 30% for TP and 46% for TN. EC remained unchanged. The presence of more anionic than cationic surfactants in the greywater is likely due to the type of household products used, such as washing detergents but no softeners. Surfactant removal was effective, achieving 88% for cationic and 96% for anionic. The BOD₅/COD ratio was 0.31 before and 0.34 after treatment, indicating limited biodegradability (Rudaru et al. 2022), suggesting that much of the removal occurred through physical removal processes. This assumption is supported by the high removal of turbidity. The pH slightly increased from 6.9 before to 7.3 after treatment, remaining within the optimal range for biological treatment, which is typically between 6 and 8 (Meenakshipriya et al. 2008). Microbial reductions in this study ranged from log 2.1 for E. coli to log 4.0 for P. aeruginosa.

Comparing these results to the pre-study (Buehler et al. 2024), the current study showed slightly better COD removal (88% vs. 85%) and consistent turbidity removal (97%). Surfactant removal was improved in this study, with 88% for cationic and 96% for anionic, compared to 79% in the pre-study, though different analytical methods were used. The pre-study had higher initial pollution levels, with COD at 506 vs. 113.4 mg/L, and turbidity at 204 FNU compared to 37.4 FNU in this study. Total aerobic count reductions improved from log 0.6 to log 3.6 in this study. Reduction of coliforms remained consistent (log 2.8). Removal rates in other aerobic attached growth systems ranged from 64 to 99.5% for COD and 53–95.6% for BOD₅ (Khalil & Liu 2021), positioning our results in the upper range. To assess the compliance of the treated greywater in KREIS-Haus with regulations and guidelines, Table 3 compares the measured values with the fit-for-purpose approach, considering specific limit values for non-potable uses. Where available, official regulations from Switzerland or the EU were used, such as the Swiss regulation for water used in public showering and bathing facilities (EDI 2016) and the EU irrigation guideline for Class A water used for food crops (European Commission 2020). For other uses, non-binding guidelines were applied, including the guideline for laundry water by the Flemish Institute for Technological Research (VITO 2020) and the guideline for toilet flushing by the German Association for Water, Wastewater and Waste (DWA 2017). The treated greywater is suitable for irrigation and laundry and potentially suitable for showering and toilet flushing, though an assessment of all stated parameters in the guidelines is advised. Regarding Swiss regulations for discharge into water bodies, the overflow tank's average COD level (50.3 mg/L) was within the limit value of 60 mg/L COD (Der Schweizerische Bundesrat 1998). TP was 0.53 mg/L and TN 3.28 mg/L. During the experimental period, the overflow tank was emptied four times. However, it is unclear if compliance with regulation can be ensured with higher occupancy of the house. Therefore, operations must continue to be monitored over a longer period with higher occupancy. Furthermore, the analysis identified two key shortcomings in the current regulation. First, there is a lack of official, binding standards for various non-potable applications. Second, evaluations based on traditional limit value approaches fail to account for potential threshold exceedances between sampling periods. In contrast, quantitative microbial risk assessment provides a more comprehensive alternative by estimating actual health risks using empirical data and probabilistic methods, which better capture variability and uncertainty in system performance and conditions (Zhiteneva et al. 2020).

However, water reuse systems are by nature vulnerable to water quality issues caused by user behaviour, such as pouring chemicals down drains or using showers as toilets. To mitigate health risks, continuous online monitoring of microbial contamination or suitable surrogate parameters is essential for ensuring safety and enabling the widespread adoption of these systems (Reynaert et al. 2024).

The measured values for the water system (Supplementary Material, Figure S3) resulted in an average daily energy consumption of 2.88 kWh, with a median of 2.78 kWh and a standard deviation of 0.77 kWh. The lowest daily consumption recorded was 0.79 kWh, while the highest reached 6.32 kWh. In comparison, calculating the energy consumption for the rainwater and greywater treatment systems separately resulted in a total daily energy consumption of 2.52 kWh (Supplementary Material, Table S5). Thus, the actual measured values were slightly higher, most likely due to longer running times of the pressure pumps on certain days with higher water usage.

The greywater treatment consumed an estimated 1.88 kWh/day to treat 0.2 m³, resulting in an energy consumption rate of 9.4 kWh/m³ or 1.88 kWh/day (Table S5). The calculation with optimization measures led to a reduced energy consumption of 2.6 kWh/m³ or 1.25 kWh/day (Supplementary Material, Table S6). Other biological on-site greywater treatment systems for reuse reported energy consumption of 0.8 kWh/m³ (Künzle et al. 2015), 1.4 kWh/m³ (Jabornig & Podmirseg 2015), and 1.79 kWh/day (Lakho et al. 2021), placing our system at the higher end of the spectrum. The rainwater treatment in KREIS-Haus had an estimated energy consumption of 6.08 kWh/m³. In comparison, other studies reported values ranging from 0.1 to 5.3 kWh/m³ (Retamal et al. 2009; Siddiqi & Fletcher 2015), but often without the inclusion of a disinfection step.

The relatively high energy consumption of the KREIS-Haus system is partly due to the use of two separate treatment systems, a dry toilet, and reusing treated greywater, which results in each system treating a smaller volume of water than in a typical household. This causes the base energy consumption to be spread across a reduced amount of water. Despite the higher energy usage, the key advantage of the system is its ability to operate fully off-grid and achieve water self-sufficiency, eliminating the need for a parallel connection to centralized infrastructure, and making direct comparisons with other systems challenging. However, there is potential to improve the energy efficiency of treatment processes. Increasing the occupancy of the building or using the system to serve multiple housing units could enhance scalability. While this would proportionally increase the energy required for pressure pumps and UV-LED lamps, it would also spread the fixed energy consumption during standby over a larger volume of treated water, improving efficiency per volume.

Furthermore, energy consumption should be assessed within the context of the fit-for-purpose approach implemented in this study. Treating water only to the necessary quality standards generally reduces energy demand, as higher water quality typically requires greater energy input (Capodaglio 2021). Without the application of the fit-for-purpose approach, the energy consumption in this system would likely have been even higher.

The sensitivity analysis showed that an increase in maintenance from 10 to 20 h strongly impacts the total life cycle cost, causing a 35% increase in the scenario of KREIS-Haus without connection fees. Conversely, and at an hourly rate of CHF 40, the analysis revealed a potential 17% reduction in total costs (Supplementary Material, Table S4), indicating a need to improve maintenance efficiency, especially in regions with higher labour costs. Implementing remote monitoring systems could offer continuous data tracking and instant alerts, potentially reducing the need for frequent on-site maintenance (Lee et al. 2008). However, this might also increase costs, depending on the availability of cost-effective and reliable monitoring solutions. Additionally, enabling homeowners to perform basic maintenance could lower costs by reducing reliance on professional services, but it could introduce new risks associated with untrained individuals managing the systems. Moreover, decentralized systems carry a higher risk of water quality issues, which could result in unforeseen costs, complicating a direct cost comparison with centralized systems. On the other hand, if more decentralized systems were implemented regionally, there could be opportunities for cost reductions through streamlined and grouped maintenance efforts.

This study evaluated the technical and economic feasibility of the KREIS-Haus off-grid water system, providing a unique demonstration of fully off-grid water self-sufficiency in Switzerland. The results contribute to addressing the identified gaps in decentralized water systems research, particularly concerning long-term operational performance under variable occupancy.

Throughout the experimental period, the KREIS-Haus achieved complete water self-sufficiency. Simulations indicated that this could be sustained with continuous occupancy of up to three people. However, variability in daily and annual rainfall, along with the impacts of climate change, such as altering rainfall patterns and longer dry periods, must be considered when interpreting and applying these results. Adaptive strategies, such as expanding the rainwater catchment area, increasing the rainwater storage tank, or increasing greywater reuse for closer-to-potable applications, could help maintain a high level of self-sufficiency.

The rainwater and greywater treatment systems demonstrated effective pollutant removal rates, with greywater treatment showing high COD and turbidity removal efficiencies. The water quality of the treated rainwater showed that most parameters met Swiss drinking water regulations, as specified in Section 3.2, though some, such as enterococci and copper, occasionally exceeded limits. This suggests the need for improvements, such as enhanced filtration, additional disinfection, and better roof maintenance. The treated greywater met the applied guidelines for irrigation and laundry use, and possibly for showering and toilet flushing, as specified in Section 3.3. However, key shortcomings of current regulation include the lack of binding standards for non-potable uses and the limitations of traditional limit value-based assessments, emphasizing the value of adopting risk-based frameworks.

Energy analysis indicated relatively high consumption compared to other decentralised water treatment systems. This is largely due to the smaller volumes treated in such an off-grid setting, where energy use is spread across lower water volumes. Despite higher energy usage, the system's off-grid functionality and water self-sufficiency offer key advantages. Optimizations, such as increasing building occupancy or serving multiple units, could improve energy efficiency by distributing base energy consumption over a larger water volume.

From an economic perspective, KREIS-Haus could potentially be competitive over a 30-year period compared to traditional water infrastructure, but only when connection fees are excluded. This highlights the impact of regulatory frameworks on the economic viability of innovative water systems. Maintenance emerged as a crucial cost driver, highlighting the importance of improving maintenance efficiency. Remote monitoring and homeowner involvement in basic maintenance could reduce costs, though both introduce risks, including potential water quality issues and additional unforeseen expenses.

The results in this study were generated under the use of the KREIS-Haus as a holiday home, with part-time occupancy and varying user behaviour. While some findings, such as water self-sufficiency and treatment efficiencies, may be transferable to a typical residential setting, differences in usage patterns, such as more consistent water input and higher occupancy, could influence system performance. Factors like increased greywater volumes, continuous rainwater demand, and user behaviour would need to be considered when adapting these results to a full-time residential context.

Future research should explore the long-term performance of the KREIS-Haus system under continuous residential occupancy to validate its scalability and efficiency. Investigating the effects of variable rainfall patterns is crucial for optimizing water self-sufficiency in diverse climatic conditions. Additionally, more comprehensive assessments of user behaviour and its impact on water quality and system performance are needed. Lastly, flexible regulatory frameworks tailored to decentralized systems and innovative continuous monitoring technologies should be prioritized to support widespread adoption and long-term sustainability.

Philippe Schläpfer, Laila Lüthi, and Tabea Vischer are acknowledged for their assistance with sampling and lab analysis. Matthias Frei and Jürg Abderhalden are acknowledged for their contributions to the design of the greywater treatment technology.

The research was funded by the Swiss Federal Office for the Environment in the scope of project no REF-1011-07300 ‘KREIS-Haus – Direktrecycling von Grauwasser’.

D.B. conceptualized the process, developed the methodology, rendered support in formal analysis, investigated the study, wrote the original draft. R.B. rendered support in formal analysis, investigated the study, wrote the original draft preparation. R.J. supervised the work, wrote and reviewed and edited the article. A.S. rendered support in funding acquisition, wrote and reviewed and edited the article. D.R. conceptualized the process, developed the methodology, supervised the work, wrote and reviewed and edited the article.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.