National Technical University of Athens (NTUA)
Investigating the integrated process of upflow anaerobic sludge blanket coupled with subsurface flow constructed wetlands for the treatment of domestic wastewater
Abstract
dc:descriptionThe treatment of domestic wastewater in rural, peri-urban, and decentralized settings remains a significant global challenge. Conventional centralized treatment plants are often impractical for these areas due to high capital costs, energy-intensive operations, and the logistical burden of extensive piping networks. Consequently, there is a pressing need for decentralized, low-maintenance, and energy-efficient alternatives that support the circular economy by transforming wastewater into a reusable resource. The technological combination of Upflow Anaerobic Sludge Blanket (UASB) reactors and Constructed Wetlands (CWs) has emerged as a promising paradigm. However, several critical research gaps hinder their optimized implementation in the context of Mediterranean region. UASB reactors typically rely on ambient temperatures. In Mediterranean climates, this often means operating under sub-mesophilic or psychrophilic conditions, which can slow microbial kinetics and reduce treatment efficiency. UASB reactors are usually implemented for industrial wastewater with increased content in organic matter and, as such, the operation with diluted streams like domestic wastewater remains a challenge. Traditionally, subsurface flow CWs rely on sand and gravel of specific composition as a filling material, which may not be locally available (especially in island regions). As such, the investigation of alternative materials that could replace sand and/or gravel is under investigation and depends on the treatment requirements. CWs are also prone to "bioclogging," reducing the system's lifespan, when the pre-treatment stage is not sufficient. As such, the investigation of integrated processes with enhanced pre-treatment processes such as the UASB process allows the long-term sustainability of the configuration. While CWs are "low energy," they are known sources of greenhouse gases (GHG) like methane (CH₄) and nitrous oxide (N₂O). Current GHG emissions assessment methodologies for CWs are numerous and in some cases controversial, making it difficult to replicate the carbon footprint assessment in similar systems and ensure that measurements taken are representative for the operation throughout a specific timeframe. The primary objective of this Ph.D. thesis is to experimentally investigate and optimize an integrated UASB-CW system for treating domestic wastewater in remote areas, focusing on low operational costs and high-quality treated effluent that is suitable for unrestricted agricultural reuse. The research is structured around three main pillars. The first pillar includes monitoring of a full-scale UASB-CW system for approximately 1,000 days to evaluate treatment efficiency under fluctuating Mediterranean temperatures and organic loading conditions. This includes assessing organic matter removal, nutrient transformation, biogas production and the evaluation of the final effluent quality reuse potential. The second pillar includes the implementation of pilot-scale Vertical Subsurface Flow (VSSF) CWs to compare conventional sand with novel engineered or recycled filling materials for the treatment of the UASB reactor effluent. The assessment is complemented by investigating the materials’ properties such as granulometry, porosity, hydraulic conductivity and adsorption capacity for pollutants like ammonium and phosphate. The third pillar includes the implementation of a robust monitoring approach to measure direct CH₄ and N₂O emissions from both full-scale and pilot-scale CWs. This pillar seeks to understand how loading rates, temperature, and configuration influence GHG emissions, ultimately assessing the system's climate impact. The primary investigation involved a full-scale system located in Antissa, Greece, which was operated for 1,009 days across ten distinct operational phases. The first stage of the system’s configuration consisted of two parallel square-base UASB reactors with a combined working volume of 41 m³, operating under ambient temperature conditions to treat domestic wastewater that had undergone preliminary screening and grit removal. These reactors had at the bottom a flow distribution system and at the top a Gas-Liquid-Solid (GLS) separator at the top for biogas collection. The anaerobic effluent was then directed towards a two-stage CW system: the first stage was a continuously fed, saturated (SAT) VSSF CW (250 m²) filled with gravel and planted with Phragmites australis, while the second stage was an intermittently fed, unsaturated (UNSAT) VSSF CW (600 m²) featuring a sand layer and four different aquatic plant species (Schoenoplectus lacustris; formerly Scirpus lacustris, Iris Pseudacorus, Typha latifolia and Juncus inflexus). This unsaturated stage was managed via batch feeding to ensure a resting period longer than 3 hours, facilitating oxygen transfer and nitrification. The integrated UASB–CW system had a maximum treatment capacity of ~100 m³/d and was monitored to remove conventional pollutants (Total suspended solids - TSS, Chemical Oxygen Demand - COD, Biochemical Oxygen Demand – BOD5, total nitrogen – TN, ammonium nitrogen - NH4-N, total phosphorus - TP, orthophosphate-P - PO4-P) as well as in terms of biogas production and the potential to reuse the reclaimed water according to the EU Regulation 2020/741. Additionally, this research included a series of Biomethane Potential (BMP) tests to evaluate the performance of the UASB biomass under controlled conditions, simulating the temperature ranges and loading regimes of the full-scale reactor. To characterize the biological health and functional distribution of the reactor, microbial diversity was assessed through 16S rRNA gene sequencing, with samples collected from different reactor zones and time points. The methodological framework was supported by comprehensive mass balance calculations for COD and long-term hydraulic monitoring of CW drain patterns to evaluate the system’s stability and the risk of media clogging.Complementing the full-scale system, a pilot-scale system was developed using four plexiglass simulation columns (0.25 m diameter) each one having different filling materials. This comparative research evaluated biochar, recycled glass, and gel beads against a control sand media for 487 days under increasing load to determine their long-term efficiency in removing conventional pollutants and their potential for reclaimed water reuse according to the EU Regulation 2020/741. The properties of these materials, including granulometry, porosity, and hydraulic conductivity, were characterized, and their adsorption capacities for ammonium and orthophosphate ions were modelled using Langmuir and Freundlich isotherms. A dedicated experimental series focused on monitoring greenhouse gas (GHG) emissions (carbon dioxide – CO₂, CH₄, and N₂O) from both the full-scale and pilot-scale CWs. Using a Fourier Transform InfraRed (FTIR) multi-gas analyzer (GASMET DX4015) and a closed-loop soil chamber system, mass fluxes were estimated to be correlated with the influent loading. The monitoring of the full-scale CWs in multiple surface points enabled the estimation of the spatial variability, while the intermittent feeding of both the full-scale and pilot-scale unsaturated VSSF CWs enabled the assessment of the temporal variability of emissions during the resting period. The estimation of the CO₂ equivalent per cubic meter of treated wastewater enabled the evaluation of the treatment footprint and the comparison with other wastewater treatment technologies. These measurements were conducted under varying operational loads and temperature. The long-term (3 year) operation of the UASB process showed that in order to sustain average TSS and COD removal efficiencies within the ranges of 57–69% and 59–76%, respectively, the organic loading rate (OLR) should be maintained between 1.16–1.50 kg COD/m³-d when the average reactor temperature is 22.7–26.4 °C. Under these conditions, high removals can be achieved through proper microbial adaptation. However, the OLR value should be reduced to 0.19–0.7 kg COD/m³-d when the temperature averages 13.5–15.5 °C, allowing the reactor to provide moderate removal (59–64%) alongside process stability. The lower performance observed under sudden OLR increases show that such abrupt changes should be avoided – even at moderately high temperatures – to ensure gradual adaptation and sufficient hydraulic retention time (HRT). Ther temperature correction factor of θ = 1.092 that was obtained suggests that temperature significantly influences the substrate utilization rate within the sub-mesophilic range of operation. Moderate total nitrogen (TN) removal, ranging approximately from 13% to 24% on average, was observed in the UASB reactor. However, some cases showed zero or negative removal values dipping below -10%. This suggests potential nitrogen release or conversion to other nitrogenous forms, likely due to biological transformations under anaerobic conditions. Given that UASB systems are not specifically designed for nitrogen removal, this modest and variable performance aligns with expectations. Similarly, total phosphorus (TP) removal exhibited variability and generally moderate values, with some periods displaying negative values that indicate phosphorus release rather than removal. Average removal efficiencies peaked between 14–25%. The overall trend suggests the UASB reactor struggles with consistent phosphorus removal, likely due to the absence of the aerobic conditions or chemical precipitation processes typically required for effective phosphorus reduction. The sludge pH in the UASB reactor remained stable within a range of 6.8 to 7.4. Soluble COD within the reactors did not significantly exceed that of the UASB effluent, and the neutral pH maintained throughout suggests that substantial volatile fatty acid (VFA) accumulation did not occur. Alkalinity was found to be higher in the UASB effluent than in the influent, an increase that suggests effective buffering due to bicarbonate formation during methanogenesis. During stable operation under low temperatures and a low OLR (13.5 ± 0.7 °C and 0.19 ± 0.07 kg COD/m³-d), methane production was 1.9 ± 0.6 Nm³/d. In contrast, under higher temperature and loading conditions (22.7 ± 3.8 °C and 1.32 ± 0.31 kg COD/m³-d), methane production rose to 12.2 ± 2.7 Nm³/d. A high CH₄ content was observed at 84.6 ± 5.0 %, while CO₂ ranged from 7–10% and oxygen O₂ remained below 1%. Other gases, such as H₂S, were below 3%. The annual average methane yield was 0.36 Nm³/kg CODrem, with a range of 0.13 to 0.92 Nm³/kg CODrem depending on influent quality and temperature variations. More specifically, the methane yield was very close to the theoretically expected value (0.35 Nm³/d), under almost mesophilic conditions (e.g., during summer period operation). On the other hand, it was much lower at low temperature and lower organic loading due to the lower activity of methanogens. When temperature and loading increased, methanogenesis recovered and, as such, the methane yield was higher than the theoretical value due to the biodegradation of previously entrapped organic matter. The BMP test results revealed that methane yields generally increased with temperature, with the highest values observed at 28.1 °C and 38.0 °C. These peaks were associated with higher food-to-microorganisms (F/M) ratios (0.74 and 0.82), indicating that the combination of sufficient substrate and elevated temperature supports optimal biomethane production. At lower temperatures – especially below 12 °C – yields remained consistently low across all F/M values, suggesting that metabolic limitations under cold conditions suppressed methane generation regardless of substrate availability. The F/M ratio was found to play a secondary role, as its influence on yield only becomes pronounced when temperatures permit active microbial conversion. Together, these results underscore the interdependence of temperature and organic loading in determining methane yield efficiency. The observed biomass yield ranged from 0.03 to 0.08 kg VS/kg CODrem. During certain periods (i.e. winter periods), an accumulation of organic matter occurred that was not degraded within the reactor. Conversely, in other periods, previously entrapped particulate COD was degraded without being registered as "removed," which explains the variation in biomass yield. At the phylum level, the microbial community was predominantly composed of Synergistota (31.8–46.1%), Bacillota (16.4–21.8%), Pseudomonadota (6.4–9.9%), and Chloroflexota (5.0–9.1%) across all samples, with Synergistota being the dominant group. Notably, differences between the top and bottom reactor sections were minimal at each time point, suggesting that temperature conditions, rather than reactor stratification, were the primary factors shaping microbial community dynamics. One specific cluster of microorganisms thrived temporarily under colder conditions: Trichococcus sp. was nearly absent at warm temperatures but increased markedly during the colder phase, reaching 2.3% in the upper zone and 1.04% in the lower zone. Trichococcus spp. are known psychrotolerant bacteria frequently detected in anaerobic digesters processing sludge. The monitoring of the CW operation shows that in order to sustain removal rates of 92–98% for TSS, 77–91% for COD, and 88–99% for ammonium nitrogen (NH₄-N), the OLR of the VSSF SAT CW and VSSF UNSAT CW should be maintained at 69.1–89.5 g COD/m²-d and 10.9–19.9 g COD/m²-d, respectively, when the average wastewater temperature is 22.7–26.4 °C. The NLR in the VSSF UNSAT CW should be 6.8–10.7 g NH₄-N/m²-d. These values should be reduced to 10.1–23.3 g COD/m²-d, 2.0–6.5 g COD/m²-d, and 1.4–3.1 g NH₄-N/m²-d, respectively, when wastewater temperatures average 13.5–15.5 °C. A low to moderate temperature dependence (θ= 1.138) was found for NH₄-N nitrification based on the simplified Arrhenius equation. A small drop in pH was observed across the treatment stages, particularly in the VSSF UNSAT CW, where biological oxidation and CO₂ production increased acidity. This trend is consistent with nitrification and the formation of organic acids. Decreased alkalinity was also observed in the VSSF UNSAT CW, with values dropping as low as 61 mg CaCO₃/L, indicating significant consumption of alkalinity during nitrification. This continuous decline in alkalinity, accompanied by a gradual reduction in the pH of the final effluent, suggests that the filter media may have contained limited buffering minerals, such as limestone, which were progressively depleted during operation, thereby reducing long-term buffering capacity.A progressive decline in the drainage performance of the VSSF UNSAT CW was observed over time (i.e., peak flowrate decrease from 0.13 to 0.10 m³/min over a two-year period which indicates lower drainage intensity and greater drainage time). The reduction in peak flow rates and total drained volume points to aging effects such as clogging, which could reduce long-term treatment efficacy if left unaddressed. Regarding the integrated UASB–CW system, overall TSS removal efficiency was very high (97–99%), achieving final effluent TSS concentrations of 2–5 mg/L. Overall COD and BOD₅ removal efficiencies were also very high, reaching 91–96% and 94–99%, respectively. The final effluent COD was mostly soluble, averaging 19–40 mg/L, while BOD₅ ranged from 3–12 mg/L during steady-state operation and 7–20 mg/L during transitional phases. The TN and TP removal never exceeded 45%, demonstrating the system's capacity to produce effluent suitable for fertigation. This work proves the feasibility of the integrated UASB–CW decentralized scheme to meet BOD₅ (10 mg/L), TSS (10 mg/L), and turbidity (5 NTU) limits for unrestricted agricultural irrigation according to EU Regulation 2020/741 Class A. The treated effluent consistently met local legislation limits for discharge into non-sensitive water bodies, with treatment and reuse costs ranging from 0.056–0.34 €/m³, considering the variable treated wastewater volume and the potential energy recovery from the UASB reactor, which also varies in different seasons.In the pilot-scale VSSF CWs, variations in the granulometric curves and porosity among the filling materials of sand, biochar, recycled glass, and hydrogels resulted in comparable hydraulic conductivity for sand and glass, while that of biochar was lower and that of hydrogels was higher. In nearly all cases, the Freundlich equation yielded higher R^2values than the Langmuir model for ammonium and orthophosphate ions, indicating that adsorption followed a heterogeneous pattern. TSS removal across sand, biochar, and glass was consistently very high (96–97%), validating the sustainability of these materials under peak SLR (14.5 ± 4.7 g TSS/m²-d). Under the highest OLR (34.1 ± 4.5 g COD/m²-d), sand and glass -filled pilots achieved 83–84% COD removal, while the biochar column performed significantly better (p < 0.05) at 91 ± 3 %. Biochar’s higher porosity (0.66) compared to sand and glass (0.38) likely allowed for greater biofilm growth and enhanced biodegradation. Regardless of the NLR or temperature, ammonium removal across all three materials was very high (96–100%). During the initial start-up phase, while ammonium removal was nearly complete, nitrate was not detected for the first 9–10 days in sand, glass, and gel beads, and for over 100 days in biochar, indicating a delay in nitrification and that ammonium was removed through adsorption. Overall, biochar and recycled glass are effective for treating UASB effluent for Class A reuse.The final effluent pH of the pilot VSSF CWs ranged from 7.3 to 8.4 across all materials. The high turbidity of the UASB effluent (62–171 NTU) was significantly reduced by all pilots, reaching 0.8–4.9 NTU for sand, biochar, and glass, and 2.9–17.6 NTU for hydrogels. Based on current market prices, recycled glass can be 3–30 times less expensive than sand, whereas biochar costs 8–60 times more. However, biochar warrants further investigation for the removal of contaminants of emerging concern (CECs), where its higher adsorption capacity may excel in long-term operation. Regarding GHG emissions from the VSSF SAT CW, higher CH₄ emissions were observed in winter period; this can be attributed to lower temperatures (14 °C) and the associated increase in CH₄ solubility in the UASB effluent. The influence of temperature is further evident when comparing summer periods, as the warmer conditions coincided with the lowest CH₄ emission values. N₂O emissions were markedly lower during winter and peaked during the warmest periods, suggesting that elevated temperatures and increased loading are associated with higher N₂O release. Observed spatial and temporal variability underscores the dynamic nature of biochemical processes in VSSF CWs, reflecting the interaction between microbial activity and hydraulics that is influenced by biofilm development, solids accumulation, and root growth. In the VSSF UNSAT CW, low CH₄ emissions across all periods support the presence of effective re-aeration and aerobic conditions. N₂O emissions varied, with longer resting periods – where the bottom 30 cm remained saturated – potentially favoring denitrification and increasing N₂O release under higher nitrogen loads. The increased oxygen availability in the unsaturated CW appears to enhance nitrification and aerobic respiration. Flow pattern variations within the filter media likely impact GHG emissions by creating new zones of microbial activity. CO₂ emissions tended to remain stable under higher loading and temperature conditions during resting periods, whereas a reduction was more likely under lower loading or temperatures. CH₄ emissions were minimal and closely tied to influent type, declining significantly during resting phases, while N₂O emissions showed a notable reduction under lower loading and longer resting periods. The CO₂,e interquartile ranges were 0.14–0.61, 0.21–0.41, and 0.30–0.78 kg CO₂,e/m³ for different reporting periods – values comparable to activated sludge systems with anaerobic stage for enhanced biological phosphorus removal. Higher CO₂,e values were linked to shifts in dominant contributors: CH₄ from the saturated CW accounted for 61.8% of CO₂,e in the summer, while N₂O from the unsaturated CW dominated during a late spring transition (62%). The unsaturated CW's contribution to the total emissions from the CWs increased across the periods ranged between 29.6% to 63.7%. When grouped by gas, N₂O accounted for 23.9% to 86.4% of total CO₂,e, indicating a shift from CH₄-dominated emissions at steady state to N₂O dominance under increasing loading and temperature conditions.In the pilot-scale VSSF CWs, CO₂ and CH₄ emissions did not show statistically significant differences among materials or conditions. However, N₂O emissions differed significantly between the two operational phases across all materials. Normalized N₂O emissions were 0.34–0.50 kg N₂O-N/kg NH₄-Nin×100% for the 3-hour resting phase and 0.07–0.15 kg N₂O-N/kg NH₄-Nin×100% for the 6-hour phase, indicating that lower loading and temperature significantly reduced emissions. During resting conditions, CO₂ emissions decreased by 32–49% after 1.5 hours in the 3-hour phase, and by up to 85% after 5 hours in the 6-hour phase. Since CH₄ emissions primarily resulted from the stripping of dissolved methane in the influent (i.e. UASB effluent), no significant differences were observed between materials over time. In the 3-hour resting period phase, N₂O emissions increased by 51–61% for biochar and glass 1 hour after the feeding, while in the 6-hour resting period phase, substantial increases (up to 148%) were observed after 1.5 hours before decreasing by the 5-hour mark. This consistent increase between 1.5 and 3 hours suggests a "lag time" for nitrogen-related biochemical processes that appears independent of the filling material.Regarding the pilot VSSF CWs, CO₂,e interquartile ranges during the 3-hour resting phase were 0.04–0.19 kg CO₂,e/m³ across materials, with medians decreasing in the 6-hour phase. Biochar reduced median CO₂,e emissions by 21–48% relative to sand, likely due to the adsorption of gaseous emissions, and showed narrower interquartile ranges. CO₂,e values from the 3-hour phase were comparable to the full-scale unsaturated CW. During the 3-hour resting phase, N₂O was the dominant contributor to CO₂,e (65–89.5%), whereas CH₄ became more significant during the 6-hour phase, ranging from 61.2% to 63.9% for biochar and sand. To conclude, this Ph.D. thesis demonstrated the long-term feasibility (over 1,000 days) of a full-scale, integrated UASB reactor and CW system for decentralized wastewater treatment in temperate climates. Despite fluctuations in UASB performance caused by seasonal temperature changes and variable influent quality, the CWs ensured stable treated effluent quality that consistently met the EU discharge standards and, under most operating conditions, the requirements for unrestricted agricultural irrigation (EU Regulation 2020/741, Class A). The system achieved high removal efficiencies for suspended solids, organic matter, and ammonium, while also producing recoverable methane, highlighting its energy recovery potential. In addition, biochar and recycled glass proved to be effective and sustainable alternatives to sand as wetland media, with comparable pollutant removal performance, lower costs (for recycled glass), and enhanced biofilm development (for biochar). Finally, greenhouse gas emissions varied with season and operating conditions, but could be significantly reduced through optimized design and operation, supporting the environmental sustainability of nature-based wastewater treatment systems.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Seintos, Taxiarchis
- Σέιντος, Ταξιάρχης
Subjects
dc:subject × 23- Τεχνητοί υγροβιότοποι
- Κατακόρυφοι τεχνητοί υγροβιότοποι υποεπιφανειακής ροής
- Αναερόβιος αντιδραστήρας ανοδικής ροής
- Αναερόβια επεξεργασία
- Αέρια του θερμοκηπίου (ΑτΘ)
- Πληρωτικά υλικά
- Constructed wetlands
- Vertical flow constructed wetlands
- UASB reactor
- Anaerobic treatment
- Greenhouse gas emissions
- Filling materials
- Περιβαλλοντική Μηχανική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη και Τεχνολογία υδάτων
- Διαχείριση και Διάθεση αποβλήτων
- Επιστήμη Πολιτικού Μηχανικού
- Πολιτική μηχανική
- Environmental Engineering
- Engineering and Technology
- Water Science and Technology
- Waste Management and Disposal
- Civil Engineering
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62616