Project topics and materials
Get Complete Project Material File(s) Now! »
Constructed wetlands
Wetlands are areas whose location in the landscape makes them either permanently or seasonally saturated with water. They stay wet long enough to exclude plants species that cannot grow in saturated soils and to alter soil properties via the series of chemical, physical, and biological changes that occur during flooding (Kadlec & Wallace 2009).
Constructed wetlands are engineered systems designed and constructed to utilize the natural processes that take place in wetland vegetation, soils and their associated microbial assemblages to assist in treating wastewater (Vymazal 2007). They are designed to exploit many of the processes that occur in natural wetlands but to do so within a more controlled environment (Hammer 1989). Although originally used to treat domestic and municipal wastewater, constructed wetlands are now employed to treat a large variety of wastewaters including stormwater runoff, industrial wastewater, agricultural sewage, leachate, and more.
According to Fonder & Headley (2013), CW can be categorized under various design parameters (Figure 15), but the two most important criteria are hydrology (flow direction, media saturation and influent loading) and vegetation (type of macrophyte growth).
Figure 15: Classification of CW for wastewater treatment according to Fonder & Headley (2013). Average removal efficiencies are lowest for surface-flow (SF) CW, mainly due to the limited contact with the soil or filter medium, whereas the best all-round performances (total nitrogen removal is poor) are recorded in subsurface-flow (SSF) CW, especially vertical flow systems (Rousseau et al. 2004).
This study is focused on SSF-CW. In agreement with hydrology-based criteria, SSF-CW are classified as :
• Horizontal-Flow Constructed Wetland (HFCW) ― this CW commonly uses a gravel filter with planted vegetation, with flows passing through the bed horizontally in a saturated medium.
• Vertical-Flow Constructed Wetland (VFCW) ― composed of a gravel or sand filter, with planted vegetation. The water is distributed homogenously onto the surface, and flows vertically in an unsaturated medium.
In France, the Cemagref [French environmental and agricultural land management research institute] developed the first filters built at Saint Bohaire (Lienard 1987). The system then expanded rapidly across the country during the 1990s, and SSF-CW have since gained a good reputation as a solution for small communities in France (Molle et al. 2005a), made popular due to its low running costs and requirements ideally geared to smaller communities where only investment costs are subsidized.
Subsurface CW
Vertical-flow
VFCW consist of a filter of porous medium through which water flows down vertically (inlets are located vertically above outlets). Most VFCW systems are set up free-draining (numbers 5 and 5a on Figure 15) and spend most of the time unsaturated (Fonder & Headley 2013). This subsurface flow system harnesses a biological aerobic treatment using the bacteria that grow onto the porous medium of the filter. The filter is generally a watertight excavation filled with different material layers (gravel or sand) at a grain-size that increases along the depth of the bed. The deepest layer of medium features a network of perforated drainage pipes ventilated to the atmosphere to promote passive aeration of the substrate (Cooper et al. 1996). In Europe they are mostly planted with reeds (phragmites australis). The system is developed to provide high levels of oxygen transfer by convection (batch feeding) and diffusion to enable full nitrification. Intermittent feedings (batches) are assured via a network of pipes with multiple outlets, either placed within the filter or distributing the flow across its upper surface.
The system is usually fed with pre-treated water -the ‘settling’- phase (Brix & Arias 2005) – with a first infiltration layer of the filter composed of sand (Brix 1994a), but in France the system has been adapted to treat raw sewage (Lienard 1987, 1990) and demonstrates good performances. The vertical filter needs a second stage to obtain good nitrification and organic matter removal rates (Molle et al. 2005a). As our study was carried out on a French system, we provide a more detailed description of this system below.
The French system is composed of two VF stages, the first stage built of three stacked cells and the second stage built of two parallel cells, as recommended by French guidelines (Molle et al. 2005a). Each cell in the first stage is fed for 3.5 days with a 7-day rest period, while the second stage is fed for 3.5 days with a 3.5-day rest period. The alternating feeding and rest periods are essential to (i) control the growth of the attached biomass on porous filter medium, (ii) minimize clogging due to mineralization of the organic deposits by suspended solids (SS) contained in the raw sewage which are retained on the surface of the primary-stage filters (Lienard 1990), and to (iii) maintain aerobic conditions within the filter.
Most filters are composed of three or four material layers. The first layer of the first stage is built of fine gravel while the first layer of the second stage is built of sand (Figure 17). These layers play a crucial physical and hydraulic role in the system, conferring the permeability needed to control both solid filtration and infiltration rate. The other layers play a transition and drainage role.
Each treatment stage has specific characteristics and plays a different role in treatment. Flow infiltrated through the porous medium of the first stage undergoes a first treatment step via SS retention at the surface and superficial infiltration layer. SS retention promotes the formation of a sludge layer at the surface that increases by about 1.5 cm per year (Molle et al. 2005a). Sludge is highly mineralized thus not fermentable. This deposit layer tends to decrease the infiltration rate of filter, allowing a long contact time between the wastewater and the aerobic biomass fixed in the non-saturated medium but at the same time diminishing oxygen transfers. In this stage, treatment mainly concerns SS and COD removal due to the rich oxygen supply of VFCW. This promotes higher carbon degradation than nitrification, however it does enable measurable total Kjeldahl nitrogen (TKN) removal (≈ 50–60%).
The second stage polishes off the carbon degradation and SS retention processes, but its primary role is to complete nitrification. Low COD concentrations mean there is more oxygen available in the medium and consequently less competition for oxygen between nitrification bacteria and carbon-degrading bacteria. The low SS content in the second stage means no sludge layer develops at the surface, so an infiltration layer with a finer grain-size is required to control infiltration. In French systems, the d10 of the sand has to be between 0.25 mm and 0.40 mm.
Batch feeding promotes extra in-filter oxygenation due to convection phenomena occurring during water displacement into the filter and gaseous diffusion from the atmosphere via the surface when dewatering occurs. VFCW thus provide greater oxygen transfer into the bed, consequently producing a nitrified (high NO3) effluent (Cooper & Green 1995). Nitration and nitrification can only proceed if the medium contains enough available oxygen. Oxygen consumption during these two steps is less than 4.3 gO2 per gram of NH3-N (Metcalf & Eddy 1998). On the other hand, VFCW do not provide suitable conditions for denitrification, which can only occur under anoxic conditions, and the complete conversion of NO3 to gaseous nitrogen forms able to escape to the atmosphere cannot be completed (Vymazal 2011).
The French system visibly requires a feed flow of at least 0.6 m³.m².h¹ to ensure satisfactory distribution for the first feeding after a rest period, when infiltration rates can be greater than 1.4 10-4m.s-¹ (Molle 2003). This feeding flow level ensures good sludge and water distribution on the filter. Iwema et al. (2005) recommend a maximum of 50 m² of surface for one feeding point to ensure good water distribution at the first stage and avoid short-circuits in porous media. Furthermore, the height of surface water produced at each batch has to be between 2 and 5 cm. Below 2 cm, equal distribution becomes difficult, while above 5 cm, infiltration rate increases and preferential flows get stronger, which decreases treatment performances. Molle et al.(2005a) studied more than 80 VFCW in France and found that an overall surface for both stages of 2 m².PE is a prerequisite in order to attain sufficient nitrification. Surface sizes greater than 2.5m².PE do not appear to improve removal rates.
Although the VFCW system achieves high organic treatment and nitrification performances, it only gives low phosphorous removal due to the need to mineralize organic matter and the fact that the media/phragmites cannot significantly adsorb phosphorus given the loads applied (Brix 1997). A study found high phosphorous removal performances on domestic wastewater by a VFCW just put into operation, but P removal fell to 20% once the mineral adsorption sites were saturated and even cases of P release produced by hydraulic overload. However, research teams are seeking out alternative materials for phosphorous sorption (Prochaska & Zouboulis 2006; Prochaska et al. 2007; Harouiya et al. 2011; Vohla et al. 2011). According to Arias & Brix (2005), under the usual design parameters for VF systems (approx. 2.5 m² P.E. and a 1 m bed depth), using sands with high Ca content could remove P for approximately 5 years, which represents high material volumes, but removal efficiency would then decrease over time until effluent concentration eventually exceeds discharge demands. Other materials, like apatite, shows good potential in terms of kinetics and saturation capacity, making it possible to use relatively small volumes of material for long-term P removal (Molle et al. 2005b). France already has a set of guidelines on P removal by apatite (Molle et al. 2011).
Horizontal flow
HFCW is the most widely used CW concept in Europe (Vymazal 2005). HFCW features biological anaerobic treatment. The design typically consists of a rectangular filter planted with common reed and lined with an impermeable membrane. The filter is completely saturated, and a water level control structure at the outlet adjusts in-bed water height to maintain saturated conditions in the filter during feedings without any surface flooding.
The system uses gabions (pea gravel) at the filter inlet and outlet to improve water distribution and recovery. Feeding can be continuous, although the distribution system needs to properly spread water across the full filter width. As this kind of filter is more sensitive to clogging, a pre-treatment is necessary to remove solids.
Figure 18: Transverse-view schematic illustration of a HFCW, adapted from Iwema et al (2005)
Filter components and their roles
Efficiently running a CW hinges on finding a balance between the different mechanisms at work in the filter (hydrodynamics, gas transfers, biological activity, reed activity, etc.) that interact with physical, biological and chemical mechanisms to maintain a balance in the system. Each component plays an important role in filter dynamics, where the most important processes CW are based on physical and microbial processes. The active reaction zone of CW is the root zone (or rhizosphere) which plays host to physicochemical and biological processes induced by the interaction of plants, microorganisms, soil and pollutants (Stottmeister et al. 2003). As shows Figure 19, system performances and reliability are dependent on a number of parameters, some of which are controlled (design, operation) and others uncontrolled (climate) or “predictable” (Molle 2012).
Figure 19: Interactivity between internal and external CW components, adapted from Molle (2012)
Hydrodynamics and gas transfers will be discussed in detail in chapter 2.4, but it is important here to briefly discuss the role of other components.
The role of plants
In the first stage of a French VFCW, the main impact of plants is mechanical, due to the dense development of stems that pierce the superficial deposit layer to create pathways from the outside air to the gravel layer below. Stems allow water to infiltrate through the hydraulically-limited deposit layer and, if there is no ponding, they allow air to transfer to the gravel layers. In winter, when reeds are harvested, the stems lose their mechanical role and the deposit layer becomes the limiting step for water and gas transfers. At this point, water flow is governed by the hydraulic conductivity of the deposit layer, water pressure shaped by ponding depth, and suction into the filter (see chapter 2.4.1). These parameters combine to ensure a stable infiltration rate. The mechanical role of plants is visible in the increasing infiltration rates with reed growth (Molle et al. 2006). Reeds also permit the filter to dewater during rest periods.
This mechanical effect of plants keeps the filter surface open to water and gas transfers and prevents clogging via organic matter mineralization. Nevertheless, as filters are fed by raw wastewater in the French system, an alternation between filters is necessary to mineralize the high organic matter loads applied on the filters.
The CW plants play other important roles :
• They add the bacterial density and diversity that give robustness to the system. The biofilm attached to the root system, dead reed material and porous media is largely responsible for the microbial processing occurring in system (Brix 1997).
• Their insulation impact, as the dead still-standing reed material shields the surface from winter frost (Brix 1994b). In summer, foliage coverage stops the filter drying out by providing shadow that helps bacteria develop and organic matter to mineralize.
• They assimilate nutrients. Also termed “phytopurifi cation”. Note that this kind of system as developed in France, cannot attribute plants a real nutrient removal role and nutrient assimilation stays too insignificant given the loads applied.
• Their oxygenation role. Although the roots do release oxygen into the rhizosphere that can influence the bacterial community (Barko et al. 1991; Sorrell & Boon 1992), the oxygen transfers involved are too insignificant given the treatment needs of the high
load applied. According to Brix (1997), the non-homogeneity of oxygen release by roots makes it difficult to calculate oxygen flux. Various authors have given flux — — assumptions ranging from 0.002 g.m ².d¹ (Brix 1990) to 5-12 g.m².d¹ (Armstrong et al. 1990). This large range of differences could be explained by the diversity in measurement techniques used and sampling seasons studied. However, these oxygenation rates should be compared to the 300 g of COD and about 150 g of oxygen necessary to nitrify KN per day and per m² on the first stage of the French system.
Microorganisms
The medium is the main substrate for bacterial growth in CW. The particle size of the media defines pore space and thus surface area for biofilm growth. Microorganisms are the main drivers of nutrient and organic matter transformation and mineralization (Stottmeister et al. 2003). Bacterial development is tends to be carried out in the top 10% of filter layers (Tietz et al. 2007). The zone where the endorhizosphere (the root interior) and the ectorhizosphere (the root surroundings) meet is known as the rhizoplane. It is in this area that the most intensive interaction between the plant and microorganisms is to be expected. Morvannou et al. defines the first 20 cm in the French system as a very biologically-active zone. This does not mean the under-layers are not important. Headley et al. highlighted that deeper filters yield better performances, especially for nitrification. Microorganisms present in the root system can be under numerous forms, as biofilm, bacterial colonies or without specific geometric form. They contribute to organic matter degradation and excrete other more biodegradable lower-molecular-weight matter. The microbial transportation processes are typically strongly dependent on water temperature, thus creating seasonal patterns of microbial transformation (Kadlec 1999).
Nitrification is the main process of nitrogen treatment in VFCW and is typically associated with the chemoautotrophic bacteria (Nitrosospira, Nitrosovibrio, Nitrosolobus, Nitrosococcus and Nitrosomonas), that derives energy from the oxidation of ammonia (and or nitrite) using carbon dioxide as carbon source to synthesize new cells. As the nitrification process is highly sensitive to physical-chemical conditions (temperature, oxygen content, carbonate content, inhibitors), it is common practice to monitor limitation conditions in aerobic processes as a process parameter. Nitrification involves two steps: one oxidizes ammonium-N to nitrite-N, in a step executed by chemolithotrophic bacteria (strictly aerobic) and nitrite-N, and the other step oxidizes nitrite-N to nitrate-N. This second step is also performed by chemolithotrophic bacteria that can use organic compounds. Thus one species of nitrite-oxidizing bacteria – Nitrobacter winogradskyi (Grant & Long 1981), cited by Vymazal (2007) – is found in both soil and freshwater.
Among classical limitations that can hinder nitrification, it seems necessary to precise the one concerning temperature, oxygen content and carbonates content.
• Temperature : The optimum temperature for nitrification ranges from 25 to 35°C in pure cultures and from 30 to 40 °C in soils. Cooper et al. (1996) put the minimum temperatures for Nitrosomonas and Nitrobacter growth at 5 and 4°C, respectively.
• Oxygen content : The process needs approximately 4.3 mg O2 per mg of ammoniacal nitrogen oxidized to nitrate-N (Metcalf & Eddy 1998).
• Carbonate content : A large amount of alkalinity is consumed, approximately 8.64 mg HCO3− per mg of ammoniacal nitrogen oxidized (Cooper et al. 1996).
Dry period/stormwater runoff treatment by VFCW
Dry periods
In dry periods, the French system demonstrates good all-round treatment performances on SS removal and COD degradation at nominal loads, with removal efficiencies of 95% and 91%, respectively, and > 85% for TKN according to Molle et al (2005a). Classical removal performances of the French system are presented in the Figures 20, 21 and 22, in relation to organic load for the first stage (100% removal represented by the dotted line).
Note that even at higher-than-nominal organic loads, COD and SS removal performances stay stable (Figure 20 and 21). For nitrification, at nominal loads, about 50%–60% of removal rates are measured on the first stage, and load level appears to impact nitrification efficiency (Figure 22).
The second stage improves SS and COD treatment performances. For COD removal, high hydraulic loads seem to decrease treatment efficiency (Figure 21). The reasons put forward to explain this negative relationship include (i) lower water retention time and (ii) low pollutant concentrations at the inlet of the second stage during rain events.
Figure 21: COD and SS treated on the second stage at different hydraulic loads (Molle et al. 2005a)
While performances appear to stay stable with organic load for COD and SS under dry weather flows, the decreasing nitrification performances with increasing nitrogen load translates a different pattern of nitrogen dynamics and transformation in the filter.
Figure 22: Treated KN on the first stage (filters with aeration drains at the bottom + intermediate aeration pipes –Dia- and classic filter –C- at diff erent temperatures with HL 0.39 m.d-1) at left (from Molle et al. (2008) and on the second stage (0.05 <HL < 2.2m.d-1) at right (adapted from Molle et al. 2005a).
As KN formed at the inlet is mainly dissolved (about 70% under NH4-N), the filtration efficiency of the filter has less impact. Consequently, nitrification performances depend on processes that mainly impact the dissolved fraction. Molle et al.(2008) and Morvannou (2012) showed that if a share of ammonium is directly nitrified, adsorption onto organic matter is a key factor. The increasing N load diminishes the share of ammonium adsorbed onto media due to saturation of adsorption sites. In addition, the increase in oxygen consumption with increasing organic load preferentially affects nitrification over COD removal. Oxygen renewal can also be reduced by a higher number of batches and greater ponding time and depth during storm events, which reduces oxygen renewal and thus nitrification. Ammonium adsorption during feeding periods is assumed to be a major factor. Loads of 10 (first stage) and 15 TKN g.m-².d-¹ (second stage) seem to be thresholds from which nitrification decreases drastically. Nitrification is also affected by batch volume (Molle et al. 2006): higher batch volumes mean greater preferential flows, thus decreasing the proportion of adsorbed ammonium.
Stormwater
Stormwater is water running off urban surfaces after rainfall over urban catchments (Marsalek J. & Chocat 2002). Urban stormwater discharges have numerous adverse effects on both urban areas (flooding) and on receiving waters (flooding, erosion, sedimentation, temperature rise and species succession, dissolved oxygen depletion, nutrient enrichment, eutrophication, toxicity, reduced biodiversity) (Marsalek, 1998). Also, particles that accumulate in the sewer system during dry weather periods get eroded throughout the rain event proportionally to the energy of the flow (Gromaire et al. 2001), thereby increasing pollutant load. Stormwater treatment is therefore vitally important, particularly as the pollutant load of metals, SS and COD carried by stormwater runoff can sometimes prove heavier than the pollutant loads coming from wastewater treatment plants. Stormwater characteristics vary strongly according to geographic area, local features, length of dry period before a rain event, intensity and duration of the rain event, local atmospheric pollution sources, topography, soil pollutant deposits, and other factors, as shown in Table 5. The net result is that urban runoff is more difficult to characterize than wastewater (Lee & Bang 2000), making the treatment of stormwater a truly complex task.
Table 5: Range of event mean SS, VSS, COD and BOD5 concentrations in different types of rain-event runoff from a French urban catchment (Le Marais, Paris) with a combined sewer system, and comparison against average rain-event concentrations quoted in the literature (Gromaire et al. (2001)
Impact of rain events on filter behaviour
Wastewater characteristics and the resulting increased flow during rain events can impact the performances and durability of the filter. Storm events will :
• increase the number of batches on the filter, thus decreasing the between-batch oxygenation period,
• influx the filter with a higher amount of solids, with a risk of clogging,
• reduce the water retention time of the constituent pollutants.
The higher number of batches produced by a storm event increases filter ponding height and ponding time. It also alters retention time due to the pressure exerted by hydraulic head variations, which plays a key role in infiltration processes (Beach et al. 2005). Consequently, the proportion of water that passes directly from filter inlet to filter outlet by preferential flows increases with hydraulic load and with batch volume, limiting ammonium adsorption efficiency. Nitrification efficiency appears to be a good parameter to evaluate the impact of rain events due the sensitivity of nitrification with adsorption rate and oxygen content that are both affected by hydraulic loads. The upshot is that good stormwater treatment performance in VFCWs is a trade-off between oxygen renewal and hydraulic load acceptance.
Studies show good treatment performances of VF with low hydraulic overloads of pre-treated water (up to a 2.6-fold dry weather load; (Avila et al. 2013). In experiments on French VF systems, Paing & Voisin (2005) observed no significant differences in effluent quality (first and second stage) for hydraulic loads varying between 0.2 and 3.2-fold the nominal capacity and for organic loads between 0.2 and 0.96-fold the nominal load. One likely explanation is that hydraulic loads lightly higher than the nominal load can improve flow distribution in the system, increasing the filter surface in contact with the pollutant without penalizing oxygen renewal and retention time, resulting in a net improvement of treatment efficiencies. According to Molle et al. (2006), no infiltration problems are observed after hydraulic overloads of 4 m.day-¹ (>10-fold the dry weather hydraulic load). Rain events can be handled while staying within the quality objectives set for the respective receiving water body (COD: 125mgL-¹, BOD5: 25 mgL-¹, KN < 10 mgL-¹).
However, while hydraulic overloads may be physically acceptable on filters, constant ponding can be unfavourable for biological activity due to a lack of oxygen renewal. The issue here is HO acceptance limits to maintain good treatment performances through to satisfactory oxygen renewal rates. The minimum in-filter oxygen transfer rate would be 28 gO2/m² day according to Platzer (1999), since it takes 4.3 gO2 to oxidize 1 g TKN. Forquet et al. (2009) affirmed that oxygen renewal in VFCW is mostly dependent on HL and number of batches per day, and showed that the threshold value (up to 170 gO2/m² day) above which a linear relationship between HL and oxygen income by convection breaks down depends on the number of batches per day, and that over this value, oxygen renewal by convection becomes limited. Consequently, oxygen renewal by convection might be a limiting factor at high HL.
This capacity to support variations in organic and hydraulic loads is a strong advantage of the VFCW, but hydraulic load acceptance limits still have to be defined
VFCW dynamics
As presented earlier, hydrodynamics greatly affects the biological processes of the system as well as in-system oxygen transfers. Here we look at filter hydrodynamics and the parameters that influence it.
Hydrodynamics on VFCWs
Water flow through porous media
A porous medium is a polyphasic environment containing air, water, medium, and biomass that involves complex nonlinear fluxes at a variety of velocities. Filter water movement is shaped by different mechanisms involving energy and water potentials where water moves from high-energy-potential points to low-energy-potential points. The intermittent feeding of VFCW complexifies the filter hydraulics, creating variably saturated conditions. For saturated conditions, gravity is the main force controlling flow, whereas in unsaturated conditions matric forces (capillarity and hydration) in addition to gravity are the principal forces controlling in-filter flow (Figure 23).
Figure 23: Schematic illustration of the main forces controlling water flow in porous media, adapted from Vincent (2011)
The interface separating two immiscible fluids is the site of energy or tension that stems from the fact that the molecules of each phase are attracted to each other via the contact surface. This tension creates superficial forces at the water/air interface, and the curvature of this interface produces a differential pressure between the liquid and gaseous phases. Differential pressure can be calculated by Laplace’s law : where R and r are the radii of curvature surface, σ is the constant of superficial tension, and Δp is the pressure differential created by the curvature of the surface.
Table of contents :
1. INTRODUCTION
2. CONSTRUCTED WETLANDS
2.1. SUBSURFACE CW
2.1.1. VERTICAL-FLOW
2.1.2. HORIZONTAL FLOW
2.2. FILTER COMPONENTS AND THEIR ROLES
2.2.1. THE ROLE OF PLANTS
2.2.2. MICROORGANISMS
2.3. DRY PERIOD/STORMWATER RUNOFF TREATMENT BY VFCW
2.3.1. DRY PERIODS
2.3.2. STORMWATER
2.3.3. IMPACT OF RAIN EVENTS ON FILTER BEHAVIOUR
2.4. VFCW DYNAMICS
2.4.1. HYDRODYNAMICS ON VFCWS
2.4.2. OXYGEN TRANSFER
2.4.3. CLOGGING
2.5. HYDRAULIC LOAD AND PERFORMANCE LIMITS
2.6. HYDRODYNAMIC MODELLING OF VFCW
2.6.1. HYDRAULIC APPROACH
2.6.2. MECHANISTIC MODELS
2.6.3. SIMPLIFIED MODELS
2.7. A SIMPLIFIED MODEL AS A DECISION-SUPPORT TOOL
3. MATERIAL AND METHODS
3.1. EXPERIMENTAL SITE
3.1.1. CHALLEX CATCHMENT
3.1.2. FULL-SCALE MONITORING
3.1.3. HYDRAULIC MONITORING
3.1.4. TREATMENT PERFORMANCES MONITORING
3.2. HYDRAULIC MODELLING
3.2.1. MODELLING OBJECTIVES
3.2.2. SIMPLIFIED MODEL
3.2.3. HYDRUS MODELLING
3.2.4. SIMPLIFIED MODEL AND FACTORS INFLUENCING HYDRAULIC ACCEPTANCE
3.2.5. SEWER SYSTEM MODELLING
3.2.6. LOCAL CONTEXT AND FILTER DESIGN
4. FILTER’S DYNAMICS
4.1. HYDRAULIC OF THE FILTER
4.1.1. TDR CAMPAIGNS
4.1.2. INFILTRATION VELOCITY
4.1.3. TRACER TESTS
4.2. SIMPLIFIED HYDRAULIC MODELLING
4.2.1. INFILTRATION CAPACITY PARAMETER AND INFLUENCING FACTORS
4.2.2. COMPARISON WITH HYDRUS
4.3. CONCLUSIONS ON THE FILTER DYNAMICS STUDY
5. FILTERS PERFORMANCES
5.1. WASTEWATER CHARACTERISTICS
5.2. TREATMENT PERFORMANCES
5.2.1. SS AND TOTAL COD
5.2.2. KN REMOVAL EFFICIENCIES
5.2.3. BOD5 AND DISSOLVED COD
5.3. GOD (GLOBAL OXYGEN DEMAND)
5.4. INTENSE TREATMENT PERFORMANCE CAMPAIGN
5.5. CONTINUOUS MONITORING BY S::CAN PROBE
5.6. ALERTS CRITERIA / BIOLOGICAL LIMITS
5.6.1. ESTABLISHING THE DYSFUNCTION ALERTS
6. LONG TERM MODELLING (RAINFALL TIME-SERIES)
6.1. THREE-COMPONENT MODEL
6.2. SENSITIVITY ANALYSIS
6.2.1. LOCAL CONTEXT INFLUENCE ON FILTER (FLOW AND PONDING)
6.2.2. PLANT INLET BYPASS LEVEL AND DISCHARGE
6.2.3. INFLUENCE OF CSO THRESHOLD ON PONDING TIME ALERTS AND BYPASS DISCHARGES
6.2.4. INFLUENCE OF FILTER SURFACE AND BYPASS HEIGHT ON PONDING TIME ALERTS AND BYPASS DISCHARGES
6.2.5. INFLUENCE OF FILTER SURFACE, BYPASS HEIGHT AND FILTER AGE ON PONDING TIME ALERTS AND BYPASS DISCHARGES
6.3. CHARACTERISTICS OF ALERTS AND DISCHARGES IN VFCW CONFIGURATIONS THAT RESPECT THE LIMITS
6.4. POSSIBLE DESIGN RECOMMENDATIONS
6.5. SHORT GUIDE FOR THE DESIGNER
7. CONCLUSIONS
7.1. BIBLIOGRAPHY
7.2. METHODS
7.3. LARGE-SCALE VFCW HYDRAULICS
7.4. TREATMENT PERFORMANCES
7.5. MODELLING RESULTS
8. PERSPECTIVES
