TREATMENT WETLANDS - CHAPTER 21 pot

735 21 Implementation of SSF Wetlands The purpose of this chapter is to provide information that will help ensure successful outcomes for projects using hori- zontal and vertical subsurface ow (SSF) wetlands (or bio- solids wetlands) by discussing the internal conguration of these systems. Common construction pitfalls and start-up errors are also discussed. Implementation of SSF wetland systems consists of the following three stages: 1. Physical design (specication of internal components) 2. Construction of the wetland 3. Start-up (commissioning) The sizing of the system can be done using different sizing tools, including loading charts, scaling factors, and rst- order modeling. The pros and cons of these methods have been discussed in Chapter 20. The rst implementation stage, physical design, involves making decisions as to the number of cells, site grading, aspect ratio, bed depth, internal piping, media size, hydrau- lics, water level control, insulation, etc. The congura- tion of these components is then typically documented in a set of technical drawings (engineering plans) and written specications. Construction of the system consists of preparing bidding documents, retaining a contractor, and physically building the wetland. This is the most critical phase of the imple- mentation because construction of SSF wetlands is essen- tially a “one-way street.” Problems with site grading and elevations become difcult and expensive to correct once the liner (if any) is in place. Problems with the liner become next to impossible to correct once the bed media is installed. Because it is so difcult to undo construction problems, the wetland designer is often retained in some advisory capacity to observe and document the construction work. The nal stage, commissioning, includes introducing the ow and pollutant loading into the wetland, as well as man- aging water levels in the wetland in a manner consistent with vegetation establishment and regulatory compliance. The overall implementation process may seem decep- tively simple, but it is not. Determining the wetland size is not the same as designing the system. Mistakes made in the technical drawings and specications will lead to serious problems down the road. Mistakes made during construction may be extremely difcult to x, and repairs are likely to cost many times more than the original com- ponents. Failure to understand how the wetland is sup- posed to operate leads to start-up problems. Any of these situations is likely to lead to the intervention of regulatory authorities. Characterization of the inuent waste stream and gen- eral site assessment issues have been previously discussed in Chapter 16. Procedures for sizing the wetland (determining the necessary wetted area) have been reviewed in Chapter 20. The next stage in the implementation process is deciding how the wetland should be congured, how internal components should be placed, and how large they are. Those issues are discussed in more detail here. 21.1 SITING General conditions of the potential site will have been con- sidered during the establishment of the basis of design. There will remain a number of possibilities for the location of the system within the overall site connes. These factors have been previously discussed in a general way in Chapters 16 and 18 for FWS wetlands; these also apply to SSF wetlands. This chapter considers additional factors that often come into play when siting SSF wetlands. These systems are typically smaller than FWS wetlands and, consequently, are often con- structed in or near existing infrastructure. CONSTRUCTION ACCESS Although the as-constructed footprint area of an SSF wet- land may be small, the designer must keep in mind that these wetland cells will be constructed with earth-moving equip- ment such as backhoe excavators. The size of this equipment requires a clear perimeter area around each wetland cell; the track size of the equipment will dene the minimum separa- tion distance between the wetland cell, the site boundary, and additional wetland cells. Large excavators are often preferred by contractors because the long reach of these machines facilitates placement of bed media during construction. An example of a construction-imposed separation distance is shown in Figure 21.1. Typical separation distances are in the range of 5–7 m around the perimeter of each wetland cell to allow access by construction equipment. Designers should keep in mind that devices such as cranes (Figure 21.2) may be required to set primary treatment devices such as settling tanks, so over- head vertical clearance (from obstacles such as power lines) may be another construction access issue. Excavations for tanks, pipes, wetland basins, and water level control struc- tures raise another access issue. Clearly, there must be an adequate location on the site to temporarily store excavated soil. With deep excavations (especially trenches), the exca- vation may have to be braced, or the sidewalls may have to be stepped back to avoid the risk of cave-in. This greatly increases the amount of site area and access needed during © 2009 by Taylor & Francis Group, LLC 736 Treatment Wetlands the construction period. This is especially important in sandy soils that have a low angle of repose and thus require large excavation widths. It should be noted that heavily wooded sites will require clearing and grubbing, as discussed in Chapter 18, and remote sites may necessitate the construction of access roads, adding to the overall construction cost. SLOPES SSF wetlands require the construction of level beds. On slop- ing sites, this will necessitate terracing of the areas where wetland beds are to be constructed. On steeply sloping sites, this will require substantial cut (earth removal) on the uphill side and substantial ll (earth placement) on the downhill side; this additional earthwork can far exceed the volume of the wetland basin itself. Consequently, steeply sloping sites are often not preferred construction areas, unless no other alternative is available. It should be noted that completely at sites may also require substantial earthworks if gravity ow through the treatment process is desired. Sites with 1–3% slopes are often the easiest to work with in terms of establishing a hydraulic prole that ows by gravity. EXISTING UTILITIES SSF wetlands are often used for homes and villages that have previously lacked wastewater treatment. Because the wetland (and any associated collection system component) is often the last utility in the ground, there is always the risk of hitting and damaging other underground utilities, such as water pipes, gas lines, electrical power lines, or telephone wires. Most utilities in North America have a “one-call” tele- phone number to dispatch an individual to locate these utili- ties. Obviously, sites with known utility conicts should be avoided, or provisions to reroute the affected utilities must be incorporated into the design process. FIGURE 21.1 Separation distance between these two vertical ow wetland cells was dictated by the need for access by construction equip- ment at The Preserve, Minnesota. The use of such a large excavator was desirable to place gravel media in the wetland cells. Alternate means of gravel placement, such as an extendable conveyor, would decrease this separation distance. FIGURE 21.2 Installation of settling tanks at The Preserve, Minnesota. Note that construction access is needed for four different compo- nents: (1) the truck and trailer used to deliver the tanks, (2) the crane used to set the tanks, (3) the excavation itself, and (4) the area needed to stockpile the excavated soil. Thus, the total area needed for construction is far greater than the tanks themselves. © 2009 by Taylor & Francis Group, LLC Implementation of SSF Wetlands 737 Biosolids wetlands are generally constructed to sup- port sludge management at existing mechanical treatment plants and are often constructed in close proximity to these treatment works. Again, knowledge of existing utilities is a critical component of the design process. FLOODPLAINS Siting of wetland systems in oodplains and oodways has already been addressed in Chapter 18. It should be noted that most ood events result in considerable sediment deposition as ood waters recede. Free water surface (FWS) wetlands are much more tolerant of this sediment load than SSF wet- lands, where the additional sediment could create potential clogging issues in the wetland bed. As a result, locating SSF wetlands in oodplains requires careful consideration of the expected frequency and severity of ooding in the project area. Biosolids wetlands will contain organic sludge with varying degrees of pathogens, depending on the age of the biosolids and level of stabilization. Because the potential to export pathogenic material exists during a ood event, a decision to locate such a facility in a oodplain would have to be carefully evaluated. REGULATORY ISSUES Regulatory requirements are often a major factor in the selec- tion of SSF technology over FWS wetlands. For instance, in the United States, approximately 25% of homes rely on on- site (single home) septic systems (U.S. Census Bureau, 1990). There are considerable health risks associated with exposure to raw sewage (or sewage exiting the septic tank). As a result, there is a large body of prescriptive codes that require the treatment and disposal of the sewage without surface expo- sure of the wastewater (U.S. EPA, 2002c). Thus, wetlands such as FWS and surface-ood vertical ow (VF) systems cannot be used in these applications, and the regulations force the designer to use HSSF wetlands or VF wetlands with buried distribution piping so the water will not be exposed. Another example of regulatory requirements that dictate the wetland selection process is airports. In cold climates, ice must be removed from the airplanes prior to take-off, and the associated de-icing chemicals generate stormwater run- off that is high in organic (glycol) contaminants. Constructed wetlands can remove this chemical oxygen demand (COD) load. However, because there is a potential for bird–aircraft strike hazard (BASH), open water bodies such as ponds and FWS wetlands are often discouraged or prohibited from being constructed near airport runways (FAA and U.S. Department of Transportation, 2004). For this reason, SSF wetlands (VF or HSSF) are preferred for treating de-icing runoff. Siting of wetlands is also inuenced by required separa- tion distances from homes, water bodies, and other landscape features. These separation distances are usually prescriptive in nature and dictated by local, state, and provincial regula- tions. In the United States, these range from approximately 3 m to over 300 m, depending on the applicable regulations and the size of the treatment system. In situations where there are limited choices for siting the wetland, these separation distances may end up not only dictating the location but also t he s hape of the wetland, as illustrated in Figure 21.3. Soil infiltration area Site boundary Dosing tank 15 m Setback 6 0 m Setback 3 0 m Setback 2 0 m HSSF Wetland Septic tanks Sewer line Roadway House pad Property line 15 m Setback 15 m Setback FIGURE 21.3 The shape and location of this HSSF wetland bed in Lake Elmo, Minnesota, were dictated by regulatory setback requirements. © 2009 by Taylor & Francis Group, LLC 738 Treatment Wetlands 21.2 LAYOUT AND CONFIGURATION Once the wetland area and location have been selected, the designer must then make a number of decisions regarding the layout and conguration of the wetland. These include The number of parallel ow paths The number and type of wetlands in each ow path Clogging dynamics Cell conguration (aspect ratio and bed depth) The hydraulic prole of the wetland system The internal conguration of the wetland cells (ow distribution, bed depth, media size and type, the need for insulating mulch, the water level con- trol structures, the liner selection and type, etc.) NUMBER OF FLOW PATHS The decision to design the wetland system with two or more parallel ow paths depends on several operational factors: The need to provide operational exibility The need to provide treatment redundancy The need to operate the system on a load-and-rest basis OPERATIONAL FLEXIBILITY When there are multiple ow paths, one treatment train can be taken out of service for maintenance and the ow diverted to the other parallel wetland cells. This exibility is a desirable attribute in a wetland treatment system, but it comes at the added cost associated with extra berms, liner, water level control structures, and earthwork. These added costs become especially signicant for very small treat- ment systems. Let us consider the example of a single home • • • • • • • • • horizontal subsurface ow (HSSF) wetland sized 300 m 2 , where a 5-m perimeter is needed for construction access. The overall site area needed is approximately 35 m r 22 m, or 770 m 2 . The results of splitting this wetland into two par- allel beds is 40 m r 22 m, or 880 m 2 . In the rst case, the wetland occupies 39% of the required site area; with the two parallel beds, the wetland occupies 34% of the site area, an inuent splitter structure is necessary, and the outlet control structure(s) must regulate two water levels. Although the split-bed approach illustrated offers more operational exibility, it comes at the cost of additional area, liner, and control structures. The added costs of these com- ponents must be weighed against the alternate methods to provide operating exibility. For instance, many single-home septic tanks provide approximately three days of hydraulic retention time, so one means of taking the system out of ser- vice for maintenance or repairs would be to simply pump the septic tank. This would give the operator about three days to complete the necessary work, and if additional time is needed, the tank could be pumped again. This alternate approach to exibility is less expensive than the split-bed sys- tem, and for this reason, single-home HSSF wetlands in the United States are commonly single-bed designs. As the size of the system increases, the added cost of parallel ow paths becomes a progressively smaller compo- nent of the overall construction cost, and providing alternate means of exibility (such as pumping of septic tanks) becomes progressively more expensive. Therefore, for larger projects, two or more parallel ow paths become the preferred design approach. For small community wastewater projects, systems with a design ow greater than 35 m 3 /d would typically have two parallel ow paths, and for larger projects, additional parallel ow paths may be warranted. For instance, the HSSF wetland used to treat aircraft de-icing runoff in Edmonton, Alberta, consists of six parallel ow paths, with two cells in each ow path (Figure 21.4). FIGURE 21.4 HSSF wetland for treatment of aircraft de-icing runoff, Edmonton, Alberta. This particular system consists of six parallel ow paths, with two treatment cells in each ow path. © 2009 by Taylor & Francis Group, LLC Implementation of SSF Wetlands 739 TREATMENT REDUNDANCY Many regulatory authorities require wastewater treatment systems to be designed to treat more than the stated design ow and load. For treatment systems in the upper midwestern United States, a commonly used redundancy requirement is that the treatment system must be able to treat 75% of the inuent ow and load, with one treatment train (ow path) out of service (Great Lakes UMRB, 1997). It should be noted that this redundancy requirement was developed for mechanical wastewater treatment plants treating raw munici- pal wastewater and may or may not be appropriate for treat- ment wetlands, depending on the type and intended function of the wetland. With a redundancy requirement, two parallel ow paths are the minimum. If the designer selects two ow paths, each wetland ow path must be designed to treat 75% of the ow and load, and the wetland must be designed to treat 150% of the design (2 r 75%). In certain cases, adding a third ow path may be advantageous. With three ow paths, two treat- ment trains will be in operation when one is out of service, so the two trains in service must each be able to treat 37.5% (one half of 75%) of the design ow and load. So, the overall wet- land is sized to treat 112.5% of the design (3 r 37.5%). The designer must then evaluate whether the savings in wetland area is sufciently justied, given the added cost of the extra berms and control structures. LOADING AND RESTING Some treatment wetlands require a loading and resting regime to function properly. For these types of systems, the loading and resting sequence will determine the number of the parallel ow paths required. Most VF wetlands are pulse-loaded at a rapid rate to ood the surface of the bed. In between pulses, water drains from the wetland and air is drawn into the wetland bed (see Chapter 2). This air movement reoxygenates the microbial biolms and maintains aerobic conditions within the wetland bed. Particulate organic matter is deposited on the surface of the bed, and microbial populations are greatest in the upper regions of the bed, where the organic loading is the highest (Langergraber et al., 2006b) As the loading period progresses, this leads to the development of a clogging mat on the surface of the bed (unless very low organic loadings are utilized). This clogging mat blocks the movement of air into the bed, promoting anaerobic conditions and ponding of the bed surface (Platzer and Mauch, 1997; Hyánková et al., 2006). Loading and resting are used as means to avoid clogging of the bed and associated surface ponding (EC/EWPCA Emergent Hydrophyte Treatment Systems Expert Contact Group and Water Research Centre, 1990; Platzer and Mauch, 1997; Molle et al., 2004a). The need for resting intervals becomes increasingly important as the organic loading on the VF wetland is increased. For instance, in France many VF wetlands are fed with raw sewage (no primary treatment); to operate successfully, three VF beds in parallel are needed in the rst stage (Molle et al., 2004a). One bed is fed with raw sewage while the other two are rested (Figure 7.31). Thus, the loading and resting operating regime of these French systems dictates that the rst stage of the treatment process must con- sist of at least three parallel ow paths. 21.3 NUMBER AND TYPE OF WETLANDS IN EACH FLOW PATH Once the number of parallel ow paths has been decided, the next question facing the wetland designer is how many cells to put in each ow path. There are several scenarios that would lead the designer to consider multiple cells in series: There is a need for greater treatment efciency (and hence greater hydraulic efciency) to increase the number of PTIS through the use of two or more cells in series. There is sufcient slope to the site that breaking the ow path into multiple cells is advantageous from a civil engineering and an earthwork perspective. There is a desire to use more than one wetland type in the overall treatment process (e.g., vertical ow to horizontal ow, or vice-versa). There is a desire for staged treatment, with dif- ferent treatment reactions occurring in different wetland cells. GREATER TREATMENT EFFICIENCY The rst case is a situation where the designer wishes to increase the overall treatment efciency of the process—or to provide the same level of treatment in a smaller area. By combining two wetland cells in series, the number of tanks in series (NTIS) is theoretically increased. The decreasing length-to-width ratio for each physical compartment means that NTIS is not doubled. The relaxed parameter, P, accounts for both hydraulic effects (N) and effects of pollutant weath- ering; so a decision to increase P must be based on the pollut- ant under consideration and the best professional judgment of the designer. In any case, the advantage of cells in series for HSSF wetlands is not large unless the design contemplates a close approach to C*. As discussed in Chapter 6, compartmentalization becomes a signicant aspect of the wetland sizing process when the goal is a very high level of treatment. The case of providing two wetland cells in series (instead of one cell) was previously explored for biochemical oxygen demand (BOD) removal in Ta bles 20.7 and 20.8. In that particular example, there was not a dramatic reduction in the required area because the design goal (94.4% reduction in mass load) allowed an efuent con- centration signicantly above (double) the assumed C* value. However, as the target efuent quality gets closer and closer to C*, the hydraulic efciency N (and hence P) becomes an extremely important part of the sizing process, because bypassing of untreated efuent in the TIS model has • • • • © 2009 by Taylor & Francis Group, LLC 740 Treatment Wetlands a large impact on the efuent quality. Here, we will consider a scenario where the wetland process must deliver an efuent quality very close to C*, where there are very large impacts of compartmentalization on the wetland size. Tables 21.1 and 21.2 consider the case of an aerated subsurface ow wetland (saturated ow) to remove COD from hydrocarbons. Because of the internal mixing induced by aeration (N  3) and pollutant weathering, the wetland is assumed to function as 2 PTIS (aerated wetlands are addressed in more detail in Chapter 24). The spatially vari- able ow P-k-C* model previously introduced in Chapters 17 and 20 is utilized here. The inuent COD is assumed to be 120 mg/L, the regulatory compliance limit is 20 mg/L, and C* is assumed to be 10 mg/L. However, because of variabil- ity in the wetland performance, an efuent quality of 11.2 mg/L is targeted as the design objective so that the system will meet the 20 mg/L regulatory limits on a consistent basis (90% of the time). The lower targeted efuent concentration is not a safety factor and has nothing to do with treatment redundancy. Further, it is assumed that there will be signicant weath- ering of the COD mixture as it approaches C*, so the two wetland cells in series are represented by P  3 instead of N  6. As the target efuent concentration is so close to C*, any bypassing of untreated water (low NTIS and therefore low PTIS) has a very large impact on treatment performance. In Table 21.1, the required wetland area is 2,400 m 2 ; in Table 21.2, the required wetland area is 1,500 m 2 . So, in this particular example, there is a large benet in terms of size reduction (35%) if the wetland is constructed as two cells in series, even though the “benet” of the second cell has been reduced by pollutant weathering effects. DIVIDING WETLAND CELLS BASED ON SLOPE On sloping sites, there will be earthwork cut-and-ll costs associated with terracing the site for wetland basins. How- ever, wetland designers should be aware that dividing the wetland into multiple segments may potentially have impacts on treatment performance (as explored in Tables 20.8, 20.9, 21.1, and 21.2). If there are large topographic changes between wetland cells, this can provide the opportunity for creative design elements, such as cascade waterfall aerators (Burka and Lawrence, 1990), if these are advantageous from a process design standpoint. MORE THAN ONE WETLAND TYPE Another obvious situation to divide the ow path into multiple cells is when there is a desire to use more than one wetland type along the ow path. This subdivision of the treatment process into unique process elements dates back to the original work of Seidel (Seidel, 1966), which is discussed in Chapter 15. Her work eventually led to the development of the current “hybrid” wetland systems, TABLE 21.1 COD Reduction for One Wetland Cell, PTIS  2 Input Parameters Calculated Values Flow rate, Q 137.1 m 3 /d Volume per tank 547.2 m 3 Precipitation, P 2.2 mm/d Area per tank 1,200 m 2 ET 0.13 mm/d Inuent ow, Q i 137.1 m 3 /d Inltration 0.01 mm/d Efuent ow, Q o 142.0 m 3 /d PTIS (system) 2 Average ow, Q avg 139.6 m 3 /d Area, A 2,400m 2 Inuent mass load 16,452 g/d Porosity, E 0.38 Efuent mass load 1,591 g/d Bed depth 1.2 m Mass percentage reduction 90.3% C i 120 mg/L Nominal HLR 0.057 m/d C* 10 mg/L HLR based on Q avg 0.058 m/d k 356 m/yr HLR based on PTIS 0.059 m/d k 0.975 m/d Nominal HRT 7.98 d Compliance C 20 mg/L HRT based on Q avg 7.84 d Variability factor 1.78 HRT based on PTIS 7.70 d Target C 11.2 mg/L Liner required 3,294 m 2 Piping required 1,200 m Calculated Values System I n Exit Tank 1 Exit Tank 2 System Out Net ow m 3 /d 137.1 139.6 142.0 142.0 Precipitation m 3 /d 5.280 2.640 2.640 — ET m 3 /d 0.312 0.156 0.156 — Inltration m 3 /d 0.024 0.012 0.012 — HLR, q m/d 0.057 0.116 0.118 0.059 Nominal HRT days 7.98 3.92 3.85 7.70 Concentration, C mg/L 120.0 21.5 11.2 11.2 © 2009 by Taylor & Francis Group, LLC Implementation of SSF Wetlands 741 which typically employ a combination of vertical–hori- zontal ow wetland cells (Urbanc-Bercic and Bulc, 1994; Cooper and Green, 1995; Platzer, 1996; O’Hogain, 2003; Obarska-Pempkowiak et al., 2004; Vymazal, 2005), or a combination of the horizontal–vertical ows (with recircu- lation) (Brix, 1998; Laber et al., 1999). Combination wet- land systems of this nature are discussed in more detail in Chapter 24. STAGED TREATMENT Another situation where a designer may elect to use multiple wetland cells along the ow path is when each wetland cell is intended to provide a different treatment function. In many situations, these wetland cells will be different types of wet- lands, as discussed previously and in Chapter 24. However, in certain applications, the same type of wetlands may be linked together solely based on treat- ment function. A good example of this approach is the two-stage VF wetland systems in France (Molle et al., 2005a). The rst stage of the treatment process is often three pulse-fed VF beds that are alternately loaded and rested. The primary function of these beds is BOD and total suspended solids (TSS) removal. The second stage of the treatment process is typically two parallel pulse-fed VF beds that are also alternately loaded and rested. These second-stage cells are intended primarily for BOD polish- ing and nitrication. 21.4 CLOGGING DYNAMICS Both horizontal and vertical ow SSF wetlands rely on the move- ment of water through a porous media for proper functioning. Both systems are susceptible to bed clogging, which restricts the ow of water through the bed media. In some instances, this restriction in ow will create ooding and will be considered to be a “failure” of the wetland treatment process. However, the denition of failure is qualitative and is often based on the per- ception and expectations of local regulators, government of- cials, system operators, and nearby residents. DEFINING FAILURE IN SSF WETLANDS Because failure is a subjective term in the treatment wetland eld, it is useful to dene what the expectations of the regula- tors, owners, and users of the system might be. Several cases concerning hydraulic failure help to illustrate this point. In Denmark, soil-based HSSF wetland technology was adopted early and spread rapidly throughout the country. Although the design intent was sub- surface ow, these systems did not have sufcient hydraulic conductivity, and overland ow occurred instead (Brix, 1998). However, the treatment per- formance of these systems was satisfactory and the overland ow mode was tolerated. In the United States, regulations for single-home septic systems prohibit the surface exposure of • • TABLE 21.2 COD Reduction for Two Wetland Cells in Series; PTIS  3 Input Parameters Calculated Values Flow rate, Q 137.1 m 3 /d Volume per tank 228 m 3 Precipitation, P 2.2 mm/d Area per tank 500 m 2 ET 0.13 mm/d Inuent ow, Q i 137.1 m 3 /d Inltration 0.01 mm/d Efuent ow, Q o 140.2 m 3 /d PTIS (system) 3 Average ow, Q avg 138.7 m 3 /d Area, A 1,500 m 2 Inuent mass load 16,452 g/d Porosity, E 0.38 Efuent mass load 1,563 g/d Bed depth 1.2m Mass % reduction 90.5% C i 120 mg/L Nominal HLR 0.091 m/d C* 10 mg/L HLR based on Q avg 0.092 m/d k 356 m/yr HLR based on PTIS 0.093 m/d k 0.975 m/d Nominal HRT 4.99 d Compliance C 20 mg/L HRT based on Q avg 4.93 d Variability factor 1.78 HRT based on PTIS 4.88 d Target C 11.2 mg/L Liner required 2,107 m 2 Piping required 750 m Calculated Values System I n Exit Tank 1 Exit Tank 2 Exit Tank 3 System Out Net ow m 3 /d 137.1 138.1 139.2 140.2 140.2 Precipitation m 3 /d 3.300 1.100 1.100 1.100 — ET m 3 /d 0.195 0.065 0.065 0.065 — Inltration m 3 /d 0.015 0.005 0.005 0.05 — HLR, q m/d 0.091 0.276 0.278 0.280 0.093 Nominal HRT days 4.99 1.65 1.64 1.63 4.88 Concentration, C mg/L 120.0 34.0 15.3 11.2 11.2 © 2009 by Taylor & Francis Group, LLC 742 Treatment Wetlands wastewater because of public health concerns (U.S. EPA, 2002c). HSSF wetlands that have clogged inlet zones and overland ow (even for a small percent- age of the wetland bed) are thus deemed “failed,” and considerable effort and expense have gone into maintaining these wetlands to eliminate overland ow, even when treatment performance was sat- isfactory. A similar situation exists in the United Kingdom (Cooper et al., 2006a). VF wetland systems in France are highly loaded with BOD and TSS (raw sewage) in the initial treatment stage (Molle et al., 2004a). A large component of the initial treatment is physical ltration and deposition of organic matter on the surface of the beds. In other words, the beds are designed to clog. However, these high loads can only be tolerated for a short time; so, a loading and resting regime (at least three parallel beds in the rst stage) is utilized so that there is suf- cient resting interval to recover from the clogged condition via drying and cracking of the cake (Figure 7.31). Therefore, in these systems, clogging does not equate to failure because there is a routine operational program to address the clogging issue. However, other VF wetlands have clogged and failed. This occurs when there are insufcient beds to allow adequate rest intervals. When these beds clog, water ponds on the surface and air movement into the bed ceases, resulting in poor treatment (Platzer and Mauch, 1997; Cooper et al., 1997). In extreme cases, ponded water may essentially ll the entire basin, requiring the entire wetland to be bypassed. Thus, it can be seen that the denition of hydraulic failure depends on the regulatory requirements, degree of treatment exibility, operational regime, and local expectations. These denitions of failure may or may not involve treatment ef- ciency issues. For design purposes in this book, the following conditions are assumed to constitute a failure mode: An inability to meet treatment standards on a con- sistent basis based on the regulatory compliance interval. For HSSF wetlands, the ow should be kept within the wetland bed. When clogging occurs in the inlet region, any overland ow path should be kept to a required minimum. This may be zero exposed water, or some minimal length that may be consid- ered an inlet distribution zone. This denes failure based on length of overland ow, not the area of the bed that is clogged. For VF wetlands, the inability of the system to hydraulically pass the design ow. This will require sufcient beds to allow adequate resting periods to resist clogging. For biosolids wetlands, not having sufcient beds to allow adequate resting periods to stabilize the applied sludge. • • • • • • CLOGGING IN HSSF WETLAND BEDS The aspect ratio (L:W) of each wetland cell is an important design decision for HSSF wetlands. Longer and narrower beds increase the organic loading applied to the cross-sectional area of the bed. High cross-sectional loadings will result in a greater length of the bed operating as an overland ow sys- tem due to the clogging mechanisms described in Figure 7.25, Chapter 7. This is likely to be unacceptable from a regulatory standpoint in the United States, and if the overland ow chan- nelizes, treatment performance will be compromised. At the present time, there is considerable interest in determining whether or not a sustainable balance between solids deposition and solids decomposition/resuspension can be reached in the inlet zone of HSSF wetlands. This issue is of profound design importance. If a steady-state criterion can be identied, then HSSF wetlands can be designed with a sufciently conservative inlet loading criteria so as to pre- clude the need for bed-media maintenance or replacement. If bed clogging is inevitable (and merely delayed by lower inlet loading rates), then the need for scheduled bed clean- ing/replacement in the inlet zone becomes an integral part of HSSF wetland technology operation and lifecycle costs. To achieve a steady state in the inlet region, the deposited inlet-suspended solids and the microbial biomat generated by particulate/soluble organic matter must be in equilibrium with the combined rate of organic solids decomposition and the rate of particulate matter resuspension. This implies that the rate of accumulation, A, in the inlet zone is zero. This con- dition can be represented (using terms previously presented in Equation 7.16, Chapter 7) as indicated in Equation 21.1: GS DR  (21.1) where generation rate, g/m ·d settling ra 2 G S   tte, g/m ·d decomposition rate of transpor 2 D  ttable solids, g/m ·d resuspension rate, g 2 R  //m ·d 2 Reduction in the hydraulic conductivity of the inlet zone can be attributed to a variety of mechanisms, including 1. Accumulation of mineral (biologically inert) sedi- ments associated with inuent TSS 2. Accumulation of particulate organic matter (sub- ject to biodegradation) within the inlet zone 3. Formation of chemical precipitates within the wet- land bed 4. Generation of microbial biolms in response to the combined load of the particulate and soluble organic matter These materials combine to form a mixture that is here termed biosolids, and the degree of biosolid accumulation depends on all four mechanisms. Because a signicant portion of the incoming organic load in domestic wastewaters is present as © 2009 by Taylor & Francis Group, LLC Implementation of SSF Wetlands 743 decomposable particulate organic matter, depositions of this particulate organic matter may result in organic loadings ten- fold higher in the inlet region as opposed to the rest of the wet- land bed (Puigagut et al., 2006). This has profound effects in biosolid formation in the inlet zone. Reductions in the hydrau- lic conductivity from biosolid accumulations are a function of the media size, as discussed in Figure 7.24, Chapter 7. Short-Term HSSF Inlet Zone Clogging At the present time, there is insufcient knowledge to quan- tify the generation G, decomposition D, and resuspension R rates within HSSF wetland beds. This is largely due to the difculties encountered when doing in situ sampling of HSSF systems, as previously discussed in Chapter 7. Without the ability to quantify the terms G, D, and R, the alternative is to extract design information from the performance of exist- in g HSSF wetlands, as shown in Figure 21.5. Data illustrated in Figure 21.5 suggests that HSSF wet- land beds should be congured such that the cross-sectional area (orthogonal to the ow) results in a BOD loading of 250 g/m 2 ·d for bed medias with a d 10  4 mm. BOD has been selected as the inuent loading parameter in this case because (1) the majority of degradable organic matter in municipal efuents is present as particulates, (2) soluble organic matter also contributes to inlet zone biosolids for- mation, and (3) chemical precipitates (as represented by the difference between COD and BOD) occur more uniformly in the HSSF bed and are likely not restricted to the inlet zone. There are drawbacks to limiting performance informa- tion to BOD, as illustrated in Figure 21.5. First of all, a HSSF wetland with a high inlet ow of nonorganic (mineral) TSS may plug at organic loads much lower than those indicated in Figure 21.5. Secondly, the nature of the organic material must be considered. Some waste streams contain organic matter that is much more degradable than others. HSSF wetlands treating soluble, rapidly degradable organic material can rea- sonably be expected to have a shorter clogged inlet zone. Lo ng-Term HSSF Inlet Zone Clogging Systems illustrated in Figure 21.5 generally have an opera- tional performance history less than ten years. Although reducing the cross-sectional organic loading may have short-term benets in reducing bed clogging, the long-term viability of HSSF hydraulic conductivity remains uncertain. Two factors argue against the ability to stabilize long-term hydraulic conductivity of HSSF beds. These include Not all organic matter deposited in the inlet zone is biodegradable. The refractory (nondegradable) component represents a long-term accumulation of organic TSS. A certain component (variable between HSSF sys- tems) of inuent TSS consists of nondegradable (mineral) material. This inert material is efciently removed from the water column through the par- ticulate settling/ltration/interception mechanisms discussed in Chapter 7. Because these removal mechanisms are very efcient in HSSF beds, most of the TSS reduction occurs in the early stages of the wetland bed (see, for example, Figure 7.26). Long-term loading of inert TSS results in an ongo- ing reduction of the hydraulic conductivity in the inlet zone. The current understanding of HSSF wetland design does not allow a quantitative determination of the region of inlet • • 0 100 200 300 400 500 600 700 800 0246810 HSSF Bed Media d 10 (mm) Cross-Sectional BOD Loading (g/m 2  d) Flooded Not Flooded FIGURE 21.5 Flooding status of HSSF wetlands as a function of cross-sectional organic loading and bed media size. (Data from Steinmann et al. (2003) Water Research 37: 2035–2042; Wallace and Knight (2006) Small-scale constructed wetland treatment systems: Feasibility, design criteria, and O&M requirements. Final Report, Project 01-CTS-5, Water Environment Research Foundation (WERF): Alexandria, Virginia; Puigagut et al. (2006) Size distribution and biodegradability properties of the organic matter in horizontal subsurface ow con- structed wetlands. Kröpfelová (Ed.), Paper presented at the 6th International Workshop on Nutrient Cycling and Retention in Natural and Constructed Wetlands, 31 May–4 June 2006; Trebon, Czech Republic.) © 2009 by Taylor & Francis Group, LLC 744 Treatment Wetlands clogging (biosolids penetration distance), although this is believed to be related to the size of the bed media, as indi- cated in Figure 21.5. The d 10 of the media is often used to approximate the pore volume, as only 10% of the bed media is smaller than this size. Studies on sand clogging indicate that the clogging dis- tance D c , can be related to the d 10 of the media: D c y 150 d 10 within the range of 5 r 10 −3 mm  d 10  3r10 −2 mm (Blaze- jewski et al., 1994). Because this d 10 is much smaller than the media sizes used in most HSSF wetlands, the usefulness of this relationship is limited to soil-based HSSF systems. Blazejewski et al. (1994) suggested a clogging thickness of 3 cm for ne-grained HSSF beds. Bavor et al. (1989) noted that clogging within a series of long, narrow HSSF trenches (L:W ratio of 25:1) was remedied by excavating the rst 5 m of the bed and replacing with coarse rock. Kadlec and Watson (1993) provided evidence from the Benton, Kentucky, system that the farthest biomat-penetration distance stretched 100 m into a 300-m bed. But this type of information does not help in design unless couched in terms of organic loadings and decomposition rates. Based on the information currently available, it appears that bed clogging (as evidenced by overland ow length) is reduced, or at least delayed, by minimizing the cross-sectional loading, and using coarser bed materials in the inlet zone. However, based on the current state of the art, there is no available evidence that argues against the phenomenon of long-term bed clogging (A  0), even if short-term clogging pitfalls are avoided. Long-term performance information from the United Kingdom (Cooper et al., 2006b) indicates that HSSF systems will require some degree of inlet zone media replacement or cleaning during the course of their operational lives. Whereas the frequency of this maintenance is dependent on the bed clogging factors described in this book, it appears that even well-designed, well-operated HSSF wetlands will require inlet zone maintenance approximately every ten years to avoid the development of large regions of overland ow. At the present time, media replacement is the most common (and most expensive) method of maintenance (Wallace and Knight, 2006) although trials using hydrogen peroxide as a cleaning agent have shown promising results (Hanson, 2002b; Behrends et al., 2006). CLOGGING IN VF WETLAND BEDS VF wetlands are also subject to solids accumulation, which can ll the pore volume within the upper region of the wet- land bed, potentially leading to ponding and hydraulic failure. Equations describing clogging of VF wetland beds have pre- viously been presented in Equations 2.58–2.61 in Chapter 2. The organic loading applied to the VF wetland appears to play a large role in the clogging process. For a soil/sand VF wetland at Merzdorf, Germany, Platzer and Mauch (1997) observed a log-linear decrease of hydraulic conductivity as a function of the cumulative applied organic load. With resting intervals, the hydraulic conductivity of the bed was restored to the original values. At the present time, there are two operating regimes to address the clogging of VF wetland beds: 1. Systems that are continuously operated without resting. These VF wetlands are loaded at relatively low organic loads to minimize biosolids accumu- lation and associated clogging of the bed. 2. Systems that employ multiple beds in a load-and- rest regime. These VF wetlands are designed to clog during the loading phase; hydraulic conduc- tivity is restored during the resting phase. Because clogging and organic loading appear to be inti- mately related, it is not possible to specify an organic load- ing rate for a VF wetland system without also specifying the resting sequence associated with that loading, as previously discussed in Chapter 20. Approximately 180 VF wetlands in North America have been designed as recirculating systems, which are function- ally very similar to recirculating gravel lters (Crites and Tchobanoglous, 1998). Recirculation ratios vary depending on design goals, but ratios of 3 to 12 times the inlet ow rate are common. An array of distribution pipes results in a nonuniform application of inuent TSS and organic matter. This loading is concentrated at the discharge orices of the distribution pipes. Biosolids then preferentially form at each discharge orice in a zone about 30 cm in diameter. In extreme condi- tions, these biosolids will plug the upper layer of the wetland bed because recirculating wetlands operate at hydraulic load- ings that are much higher than the inlet ow reductions in bed porosity and loss of hydraulic conductivity can rapidly lead to ooding of the wetland bed. An example is shown in F i gure 21.6. Operational experience in the United States indicates that the organic loading (as represented by BOD) should be less than 15 g/d per orice for buried pipes to avoid biosolids clogging with bed media greater than 4 mm. The equivalent BOD loading in the biosolids clogging region at the orice is approximately 200 g/m 2 ·d. The uniformity of application is increased and the organic loading is reduced if there is a secondary redistri- bution of the inuent by splashing or spraying. Figure 21.7 shows a cold-climate method of achieving this. Perforated pipes are placed in inltration chambers that are then buried in the upper portion of the wetland. Water sprays out of the distribution orices up into the inltration chamber, where it splashes around and undergoes secondary distribution. Ori- ce shields (resembling plastic pipe caps) are commercially available for this purpose. 21.5 CELL CONFIGURATION As discussed in the previous section, clogging dynamics will often determine the number of wetland treatment cells and, for HSSF wetlands, the aspect ratio (L:W) will be a function of the cross-sectional loading if delaying or minimizing the © 2009 by Taylor & Francis Group, LLC [...]... establishment in a treatment wetland to be successful, the plant material (seed, tuber, rhizome, pot, etc.) must Implementation of SSF Wetlands 765 Potted plant Firmly packed mulch mound to protect root mass Mulch Water level HSSF Media Water level Liner FIGURE 21. 29 Potted-plant installation procedure for a mulch-insulated HSSF wetland (From Wallace and Knight (2006) Small-scale constructed wetland treatment. .. Figure 21. 21 In other applications, it may be desirable to provide more protection to the liner This is commonly done through the use of rip-rap, as indicated in Figure 21. 22 FIGURE 21. 20 Example of poor berm finishing, liner installation, and site grading Stormwater runoff is carrying sediment into the HSSF wetland basin © 2009 by Taylor & Francis Group, LLC 760 Treatment Wetlands FIGURE 21. 21 VF wetland... indicated in Figure 23.11, Chapter 23, or rotating elbows are commonly used Water level control is essential to the establishment of plants in HSSF wetlands The minimum-cost alternative to Figure 23.11, Chapter 23, for small-scale wetlands is a fixed standpipe with removable extensions, as implied in Figure 21. 18 For large-scale systems, adjustable weir Implementation of SSF Wetlands 757 Height adjustment... strength; Figure 21. 25 is an example of this type of sidewall construction The third alternative is cast-in-place concrete walls (see Figure 1.15, Chapter 1) This is the most expensive method of construction, but offers FIGURE 21. 22 HSSF wetland divider berm armored with rip-rap, Albuquerque, New Mexico © 2009 by Taylor & Francis Group, LLC Implementation of SSF Wetlands 761 FIGURE 21. 23 HSSF wetland... flow, collect samples, or conduct bed maintenance activities (Figure 21. 14) VF WETLANDS Most VF wetlands are pulse-fed The design goal of pulse feeding is to rapidly apply the influent flow such that the bed is rapidly flooded and that uniformity of application is ensured (see Figure 2.33, Chapter 2) Typically, pulse feeding 754 Treatment Wetlands Design this interface so BOD loading does not exceed d10... requirements Final Report, Project 01-CTS-5, Water Environment Research Foundation (WERF): Alexandria, Virginia Reprinted with permission.) better control of the flow distribution (Brix, 1998) A schematic of this is shown in Figure 21. 8 For small-scale wetlands, flow-splitting approaches use weirs, stoplogs, or rotating elbows if flow balancing is an operations goal All such flow-splitting devices require periodic... techniques To date, virtually all HSSF wetlands have been designed with beds between 30 and 60 cm deep But, controlled field-scale research studies indicate that shallower depths provide better treatment VF and biosolids wetlands are typically constructed to a prespecified bed depth, 50–80 cm for VF wetlands and 50–60 cm for biosolids wetlands Implementation of SSF Wetlands 747 lower organic loadings,... gravel 748 Treatment Wetlands TABLE 21. 4 Summary of Bed Materials Used in VF Wetlands Source Wetland Type Layer Thickness Media Size Material EC/EWPCA (1990 Pulse-fed Surface layer Second layer Third layer Drainage layer 8 cm 15 cm 10 cm 15 cm — 6 mm 12 mm 30–60 mm Sharp sand Pea gravel Round gravel Round gravel ATV (1998 Pulse-fed Main layer 0.2–1.0 mm Soil/sand Molle et al (2004a) Pulse-fed (First... 60-cm bed depth, this gives h o 50 cm and hi 55 cm The two FIGURE 21. 11 Lowering of the outlet control to prevent inlet zone flooding In extreme cases, this can exceed the hydrologic tolerances of the wetland vegetation (see Figure 21. 12) (From Wallace and Knight (2006) Small-scale constructed wetland treatment systems: Feasibility, design criteria, and O&M requirements Final Report, Project 01-CTS-5,... been discussed in Chapter 18 for FWS wetlands; the same considerations apply also to SSF wetlands In North America, one-piece factory-seamed PVC liners (0.76 mm) are commonly used on small wetland basins ( 1,000 m 2) Larger projects typically use HDPE liners in the 1.1–1.5 mm thickness range; these liners are shipped to the project site in rolls Individual liner panels are fusion-welded together after . Francis Group, LLC 746 Treatment Wetlands better control of the ow distribution (Brix, 1998). A sche- matic of this is shown in Figure 21. 8. For small-scale wetlands, ow-splitting approaches. assumed to function as 2 PTIS (aerated wetlands are addressed in more detail in Chapter 24). The spatially vari- able ow P-k-C* model previously introduced in Chapters 17 and 20 is utilized here load-and-rest opera- tional strategy, as discussed in Chapter 20. Biosolids wetlands are typically governed by the rate of organic matter application (kg/m 2 ·yr), not the rate of hydrau- lic