767 22 Management, Operations, and Maintenance As discussed in the preceding chapters, most treatment wetlands are designed and constructed to require infrequent operational control or maintenance. Through conservative design and simple, low-maintenance mechanical controls, many free-water surface (FWS) wetland treatment systems will expe- rience minimal ecological changes and will continue to meet nal efuent limits for a long period of time. The situation is more problematic for horizontal subsurface ow (HSSF) wet- lands, which will require periodic bed maintenance throughout the lifetime of the system, and for vertical ow (VF) wetlands, which are dependent on loading-and-resting regimes to main- tain hydraulic conductivity, unless lightly loaded. Monitoring and adjustment of ows, water levels, water quality, and biological parameters are the principal day-to-day activities required to achieve successful performance of these low-tech treatment wetlands. Other operations and mainte- nance activities in treatment wetlands, such as repair of pumps, levees, and water control structures; vegetation management; pest control; and removal of accumulated mineral solids, typi- cally must be attended to at much less frequent intervals. This chapter provides guidance on the appropriate level of operational monitoring necessary for the control of wet- land treatment systems. Typically, these are natural systems with low hydraulic and constituent loadings. Ecological integrity therefore requires minimal attention, unless a par- ticular preconceived plant mix has been mandated. Hydrau- lic balance and hydropattern control may require infrequent attention. Compliance monitoring is driven by the permit conditions, but often a modest amount of extra monitoring is useful in optimizing operations. Emphasis should be placed on the avoidance of unnecessary operational and mainte- nance activities, including the management of nuisance spe- cies when this achieves no measurable compliance goals. 22.1 START-UP It should be fairly obvious that a newly built wetland is not in a condition to perform the functions that will ultimately prevail in the system. Yet, there have been many misunder- standings resulting from start-up conditions that do not meet long-term expectations. Start-up transients are not necessar- ily brief, because it can take considerable time for the wetland biology to develop. Several processes occur during start-up, importantly including: Plants increase in density and areal coverage. Senesced plant parts form a litter layer. • • Newly placed soils either release constituents if initially loaded or absorb constituents until they are fully loaded. Because plants form part of the “biomachine” that accretes phosphorus and other chemicals, and because plant detritus forms the sites for periphyton and bacterial assemblages, it is expected that removal will be enhanced by grow-in. One pat- tern of start-up is typied by the rapid vegetation of a bare soil wetland after an initial planting. If that planting is sparse, the start-up period will be characterized by the time for plant ll- in plus the time for litter development. That can be relatively brief, especially in warm climates. At the Orlando Easterly Wetland in Florida, the start-up duration was approximately 24 months, during which considerable change in the ecosys- tem took place. Figure 10.21 shows two of the principal vari- ables: vegetation density and phosphorus rate constant. Note the close parallel between the amount of vegetation and the rate constant during the rst two years. After that period, the vegetation density levels off, and the rate constant drops to its long-term average value of 13.5 m/yr. The phosphorus requirement for building the new standing crop of biomass leads to a peak uptake rate constant, that is, roughly double the long-term average value. An interesting feature of the unvegetated wetlands is the absence of a carbon source to fuel the conventional hetero- trophic denitrication. That process would conventionally require a supply of carbon of approximately equal mass to that of nitrogen lost (Ingersoll and Baker, 1998). A newly constructed wetland does not have the carbon source cre- ated by litter decomposition, unless the system is mulched initially with carbonaceous material. Algae are an opportunistic rst-colonizer in new wet- lands, because the canopy is not yet extensive, and sunlight is available, perhaps accompanied by copious nutrient supplies (s ee Figure 3.1). Although this initial vegetative cover devel- ops rapidly, it utilizes only a small fraction of the nutrients in the system and temporarily creates a pond environment. ANTECEDENT CONDITIONS The initial soils and sediments may contain extra nutrients or contaminants, or they may possess an initially high sorp- tion capacity. The antecedent load of pollutants should be assessed to determine if there is treatment liability that must be accommodated before the incoming water can be treated. The same processes that will eventually provide treatment • © 2009 by Taylor & Francis Group, LLC 768 Treatment Wetlands FIGURE 22.1 Start-up of a FWS wetland near Orlando, Florida. The antecedent soils were laden with labile phosphorus, which was released to overlying water during start-up. This initial load was absorbed and ushed over the rst 200 days of operation. 0 200 400 600 800 1,000 1,200 1,400 0 50 100 150 200 250 Time (days) TP (µg/L) In Out Linear (In) Linear (Out) for incoming water can also remove and immobilize such antecedent pollutants. However, if the water ow is immedi- ately commenced, then the wetland can temporarily serve as a “source” of contaminants. If unanticipated, such releases can create regulatory problems. For instance, a wetland treat- ment system in Orlando, Florida, released phosphorus and nitrogen from antecedent land uses and conditions, during a start-up in the ow through mode (Figure 22.1). Despite the improvements to performance that ultimately resulted in phosphorus removal, the regulatory expectation was that removal should occur “from day one.” Very large nes were levied just as the wetland came out of start-up, causing clo- sure of the facility. The keys to avoiding such start-up problems are an understanding of potential treatment liability and a plan to avoid discharges until the wetland has “settled down” to a post-start-up condition. However, it is difcult to forecast the direction or magnitude of potential start-up releases. Accord- ingly, a regulatory framework has been developed in South Florida for start-up of FWS wetland systems. Three periods in the (early) life of a treatment wetland are identied: Start-up phase. The permittee shall manage water depths in the treatment cells to facilitate the recruitment of marsh vegetation in accordance with an Operations Plan, which may include recir- culating waters within the system. The start-up test for an individual ow-way begins when the previously mentioned samples demonstrate, over a four-week period, a net reduction in the tar- get pollutant (phosphorus for the STAs) occurs. Discharge/ow through operations, from an indi- vidual ow-way that has passed the start-up test, may commence once documentation and all sup- porting data and analyses are submitted. The intent of this phase is to prevent ow through operation until it can be demonstrated that the project is pro- ducing net benet. • Stabilization phase. Once ow through discharges from a ow-way begin, the permittee shall initiate water quality monitoring for that ow-way consis- tent with the monitoring program. Performance during this period is not expected to be optimal because the wetland is not yet fully developed. The stabilization test is met when the long-term criteria for the facility are met, i.e., the concentra- tion limits are being met. If, after the two years of full ow through operation, the facility has not met this stabilization test, the permittee is to sub- mit a report evaluating reasons for not meeting the stabilization test. Routine operations phase. During the routine operations phase, the wetland is deemed in com- pliance if the permit efuent limitation is being achieved. In general, the South Florida projects have achieved start- up and stabilization in periods ranging from a few weeks to more than two years, which is typical of warm climates. In cold climates, a grow-in period of approximately two grow- ing seasons may be anticipated, depending of the planting d e nsity and rate of vegetation propagation (see Figure 3.12). It is generally a bad idea to initiate the development of a wetland with large ows of high-strength wastewater. Newly planted vegetation is susceptible to destruction if exposed to high-oxygen stress or high concentrations of some chemicals such as sulfur compounds. Therefore, a “ramp up” from fairly good quality water to the strong inuent is often used. For instance, the Saginaw, Michigan landll leachate treatment wetland anticipated source water of several hundred mg/L of ammonia nitrogen. The wetland was developed, including litter and plants, using rainwater and a potable water supple- ments. The system was placed on large recycle and the introduction of leachate commenced. The raw leachate was thus diluted with treated water. This start-up strat- egy was successful. As a counterexample, the potato water • • © 2009 by Taylor & Francis Group, LLC Management, Operations, and Maintenance 769 treatment wetland at Connell, Washington, was started-up with an abrupt transition from clean irrigation water to full- strength process water (anaerobic). Large populations of purple sulfur bacteria developed, and the wetland developed suldes in excess of 40 mg/L (Burgoon et al., 1999). Very strong odors were generated and inlet vegetation suffered. These were fortunately only start-up problems, which have not continued beyond the commission phase. VEGETATION START-UP Design specications for vegetation establishment in con- structed wetlands should clearly delegate responsibility for plant maintenance from the time of planting until system start-up. This task may be the responsibility of the contrac- tor or planting subcontractor, the engineer, the owner, or some combination of these parties and should be carefully described in the planting specications (see Table 22.1). The key requirements for healthy wetland plant propagules to succeed are water, soil, nutrients, and light. The rst two components must be controlled to some extent by the engi- neer and the contractor, and nature generally provides the other requirements for plant growth. The establishment of vegetation is usually a task of a planting contractor and not the facility operator. Water level control is of primary importance during the plant establishment phase. Therefore, clear lines of responsi- bility need to be spelled out: who is required to supply start-up irrigation water, and who is responsible for watering the new plants? Failure to supply water in appropriate amounts can result in planting failure (see Figure 18.35). Generally, there are three phases to water level adjustment during this phase: Phase 1: Decreasing the water level to expose the rooting media (mud at) for the initial planting or for replanting. Phase 2: A gradual increase of the water depth as emergent plants become established. The emergent • • TABLE 22.1 Example of Contractual Responsibilities for Plant Establishment 3.2 Maintenance of Stand A. Maintenance The planting contractor shall make arrangements and bear all costs of providing adequate water for initial planting and plant maintenance. Treatment efuent from the existing facultative lagoon can be used, if desired by the contractor, as long as off-site runoff is controlled. The planting contractor shall be responsible for maintaining the plants to ensure optimum growing conditions until accepted by the Owner. Initial water depth in the wetlands shall be increased gradually from 0 inches to maximum depth in 2-inch increments over a period of two months following completion of planting. All planting shall be done in moist soil without standing water. Maintenance may include watering or dewatering of the wetlands cells to ensure optimum growing conditions. B. Inspection Sixty days after the completion of the planting, the Owner or Engineer will make an inspection to determine if satisfactory stands of cattail and bulrush, have been produced. A satisfactory stand is dened as one in which: (1) spacing between planted seedlings averages 3 feet or less; (2) the seedling survival averages at least 80 percent; and (3) no areas of greater than 100 contiguous square feet with a seedling survival rate of less than 50 percent. If satisfactory stands have not been established, another inspection will be made after the planting contractor has corrected any deciencies and provided the Owner or Engineer with written notice that the stands are ready for inspection. shoots should be above the water surface at all times to give the plants access to sunlight and oxy- gen. Water level adjustments should be less than 2 cm/d during this time. Typically, the higher the oxygen stress on the plants (high organic and/or nutrient loads), the smaller the incremental adjust- ment should be. Phase 3: Setting the water level at the design depth after plants are established. Basic procedures for establishment of wetland plants for SSF wetlands have already been provided in Chapter 21. If there is signicant failure during the initial planting, these steps may need to be repeated. Initial plant inspec- tion should examine the viability of the planted propagules, whether seeds, seedlings, or eld-harvested mature plants. If plants were planted in rows on specied centers, the sur- viving plants in a subset of these rows can be counted to determine the survival rate. When counting surviving plants, each of the original seedlings or plant clumps should only be counted once, and any daughter plants that may have arisen from rhizomes should not be counted. If seeds were used, then random plots can be used to estimate germination suc- cess. Planting success can be compared to planting specica- tions based on these plant counts. In some cases, a contractor may have to ll in some areas with new plants to comply with the requirements of the planting specications. Inexperienced contractors and inexperienced engineers writing specications and providing construction supervi- sion have often killed their wetland plants through insuf- cient soil moisture, excessive water depths, inadequate soil preparation, damaged plant material, inadequate plant spacing, inappropriate planting methods, and bad timing. Success is highly probable as long as a contractor under- stands the growth requirements of the desired wetland plants and uses information available to most horticulturists. Once emergent plants are well established, replanting is usually not necessary unless the system is overwhelmed by • © 2009 by Taylor & Francis Group, LLC 770 Treatment Wetlands rodents or waterfowl, which can destroy the emergent plant community if left unchecked. In cold climates, grazing pres- sure in the winter and early spring is a signicant operational concern during the plant establishment phase, and wetland designers should anticipate replanting and/or exclusion activ- ities (Wallace et al., 2001). 22.2 MONITORING All wetland treatment systems should be monitored for at least inow and outow water quality, water levels, and indi- cators of biological condition. Following proper design and construction, monitoring is the next most important factor in successful operation of treatment wetlands. Monitoring information is essential to the ability to anticipate the need for operational changes and should be collected accurately and consistently, and frequently reviewed by a knowledgeable operator. Changing the character or function of the wetland ecosystem takes considerable time and operator investment. The appearance of the wetland is not necessarily representa- tive of what it is doing for water quality, because many wet- land functions do not tie strongly to the abovewater canopy. Consequently, it is water quality and quantity data that are the primary indicators of wetland operational status. There are few controls for Type A wetlands (the ones that are pas- sively designed) but more for Type B systems (the ones with added controllable features). Water level control is possible for all types. Added possibilities may include ow control (ll and drain), recycle, bypass, added streams with supple- mental constituents, and aeration, as discussed in Chapter 24. Alteration of the ecosystem is also possible for all types, but usually is a drastic step, to be avoided if possible. Dredging of sediments or soils, bed media replacement, new plantings, or removal of plants by mechanical or chemical means can all be accomplished but at considerable expenditure of time, money, and loss of treatment potential. In all cases, appro- priate diagnostic information is a prerequisite to initiating maintenance or control steps. W a ter Quality Continuous ow wetlands will require compliance monitor- ing of a specied list of chemical constituents and inow rate, because most operate under some kind of discharge per- mit. The sampling locations can vary from project to project, and usually include some or all of the wetland inow and outow points. However, some projects may require inter- nal monitoring points in the treatment wetland, and in some cases, monitoring of the receiving water body. The locations, types, and frequency of compliance sampling are spelled out in the individual permits, as are the averaging periods and limits for the various regulated constituents. Laboratory pro- tocols for chemical analysis are also ofcially prescribed and include a variety of established methods, as dened in Stan- dard Methods (APHA, 2005) or by U.S. EPA, for example. However, the methods for acquiring wetland water samples are neither well dened nor standard. There is no apparent difculty with obtaining samples at inlet and outlet structures, yet there can be serious anomalies related to sam- pling methodology. For instance, either autosamplers or grab sampling may be employed. The autosampler can be set to ow-proportion, and thus (hopefully) produce ow-weighted concentrations for use in mass balances. It is instructive to see what happens when both procedures are used over a long period of record (Figure 22.2). Not only do the sampling methods yield much different results, there is an apparent bias of the autosamplers to give higher results. Speculatively, the reason is that autosampler intake tubing may “suck mud” if the intake strays from its optimal position. Acquiring representative internal wetland water samples in a FWS environment is very difcult. Floating materials, such as Lemna, can often be excluded by careful sampling. Water depths may be quite shallow, and there can be large amounts of readily suspended materials in the vicinity (see Figure 7.9). The inclusion of particulates in a dipped sample will not be representative of the particulates that move with water in the undisturbed environment. The local velocities created by the sampling activity resuspend an abnormally high amount of total suspended solids (TSS). That solid material can bias the lab analyses for total amounts of con- taminants such as total phosphorus or total Kjeldahl nitro- gen (TKN). The intense vertical gradients in water quality near the wetland bottom can also create anomalous results. In very shallow water, the sample can contain some or all of pore water, which usually has much different (higher) con- centrations from surface waters (see Figure 10.5, for exam- ple). All of these factors create a need to avoid disturbance in the act of sampling and to avoid very shallow depths that may involve pore water. In the various Everglades sampling 0 100 200 300 400 500 0 100 200 300 400 500 Grab Samples Autosampler FIGURE 22.2 Autosampler versus grab samples for total phos- phorus for one of the inlet structures of STA2 of the Everglades Protection project. Data span eight years of operation. Note that the autosampler results are generally higher (mean = 223 µgP/L) than grab sample results (mean = 149 µgP/L). © 2009 by Taylor & Francis Group, LLC Management, Operations, and Maintenance 771 activities, depths less than 20 cm are not bottle-dip sampled for these reasons. However, carefully aspirated samples may be acquired at smaller depths. Similar problems exist in sampling SSF wetlands, espe- cially because dedicated sample ports must be used to extract in situ samples, which may or may not representative of the ow path. Figure 7.1 illustrates a typical sample port assem- bly for SSF wetland systems. Peristaltic pumps are often a preferred sampling device for SSF systems, as shown in Figure 22.3. Considerable care must be taken in the placement of the sampler tubing to obtain a representative sample; poor tubing placement will disturb biolms and/or suck sludge from the bottom of con- trol structures or sample ports. Dedicated sample tubing that is permanently installed is often used at facilities where rou- tine sampling is conducted. SAMPLING AND ANALYSES IN SUPPORT OF MASS BALANCES Compliance monitoring may not be enough to gain an under- standing of wetland function, because it is focused upon measuring the variables that may affect the receiving waters, and is commonly restricted to outows. A fuller understand- ing of wetland performance rests upon constituent mass bal- ances, which in turn depend upon the water mass balance for the wetland. Water Flows The water budget has been described in detail in Chapter 2. The principal inow is usually the water to be treated, which is augmented by rainfall, evapotranspiration, and seepage before forming the wetland outow. Rainfall records may be kept at the site or procured from a nearby weather station. Evapotranspiration may be derived from weather bureau pan evaporation or from governmental irrigation services, as discussed in Chapter 4. Therefore, the primary monitoring requirements are usually the wetland inow rate and outow rate. Both are required if there is signicant seepage, so that the seepage may be estimated by difference. Flow measurement can be accomplished using any one of several techniques. For pumped inows, either pump- hours plus the pump-discharge curve may be used or a tur- bine meter can be installed. For small ows, a calibrated weir, possibly with continuously recording head, is a com- mo n choice (Figure 22.4). Most large inlet and outlet devices that operate with a measurable head loss can be calibrated to determine ow from headwater and tailwater elevations. For large inows without a calibratable structure, devices such as Doppler ow meters or ultrasonic velocity meters have been employed. It is noteworthy that virtually every type of FIGURE 22.3 Efuent grab sampling using a portable peristal- tic pump at a HSSF wetland in Scandia, Minnesota. Considerable care must be taken with the sample tubing to obtain representative samples. FIGURE 22.4 A V-notch weir, with recorded head, is a user-friendly method of measuring small inows and outows. Photo from the Duck Spring wetland system near Alcoa, Tennessee. © 2009 by Taylor & Francis Group, LLC 772 Treatment Wetlands ow measurement relies upon clean conditions and periodic recalibration. Any of the measurements are prone to being somewhat in error with eld accuracies on the order of ±15%. For any of the devices that include sensors and data logging, frequent manual cross-checks are recommended. Low ows in small systems will be zero (no ow). This has several design implications. Most notably, there will be an insufcient ow to create scouring velocities in household plumbing and gravity sewer pipes. During intermittent ow, solids will “hop” down the pipe, moving only when the ow and depth exceed certain limits. Solid particles accumulate into dams, partially obstructing the invert of the pipe. When ow pulses occur, water accumulates behind the dam and pushes the solids along the pipe (Littlewood and Butler, 2003). This “sliding dam” mechanism of movement is the dominant form of solids transport in small treatment systems. Conventional weirs or upow-splitting devices are gener- ally not used as primary ow control elements for small wet- land systems unless the solid material has been removed in a primary treatment process. Due to the low ow, toilet paper and other solids will inevitably get caught on one of the over- ow weirs because there is insufcient velocity to remove the material. All ow will go to the other outlet, until it, too, clogs with solid material. Flow will switch back and forth between the outlets in an unpredictable manner and create problems for the operations staff. Accurate ow measurement for a small collection network is difcult to obtain for many reasons. Use of an overow weir as the primary control element (combined with an ultrasonic head or other sensing device) can present an ongoing problem for maintenance personnel due to solids accumulation. Mag- netic ow meters have a minimum ow velocity threshold of about 6 cm/s for accurate ow determination. For small-scale applications, this generally requires “necking down” the pip- ing to 25 or 50 mm to get the necessary low-ow accuracy. This is generally not a viable option unless solids have been removed from the wastewater in a settling tank. Small umes, such as H or HS types (as shown in F i gure 22.5), offer acceptable accuracy at low-ow rates (Grant and Dawson, 1997) but may not achieve scour veloc- ity, leaving the umes susceptible to solids accumulation. If a ume is used after a solids removal device (such as a settling tank), the velocity at low ow will allow the growth of a bio- lm on the ume surface. For this reason, bubbler units are not recommended as a sensing device because biolm growth will rapidly plug the air tube. Noncontact devices, such as ultrasonic heads, are recommended for use with umes as shown in Figure 22.5. Tipping buckets offer very accurate low-ow measurements, but the corrosive atmosphere pres- ent in sewage systems precludes the use of mechanical coun- ters. Also, tipping buckets can become “pinned” under high ows, so it is important to determine the suitability of the tipping bucket assembly at peak ow. Due to the reasons cited in the preceding text, elapsed time meters are often used as a means of ow measurement on systems that receive a pumped inuent, even though this may result in lower accuracy ow determinations. In pumped systems, the ow will be stagnant (zero veloc- ity) in pressure mains most of the time. To prevent freezing problems in cold climates, pressure mains must be installed below the frost line. If this is not possible, the pipe should be sloped to drain and provisions for a drainage outlet should be made at the pump end as well as at the discharge point. In the long term, water storage changes in the wetland are almost always negligible in comparison to other ows. How- ever, water depth is a very useful parameter for other purposes than the water budget, and hence water stage in each cell is a desirable data set. Staff gages should be installed in each cell of the wetland and surveyed for vertical control at the time of start-up (Figure 22.6). When combined with a top- ographic survey of the wetland, stage measurements provide a quantitative tool for assessing the average, maximum, and minimum water depths in the wetland and the frequency with which these depths occur. These data are essential for FIGURE 22.5 This small HS ume is used in conjunction with an ultrasonic head to record inuent ows for a HSSF wetland in Delft, Minnesota. © 2009 by Taylor & Francis Group, LLC Management, Operations, and Maintenance 773 interpreting tracer measurements of hydraulic residence time and for assessing any detrimental hydroperiod effects on biota. A variety of water level measuring instruments (e.g., analog chart recorders, ultrasonic, pressure-sensor, and capacitance- sensor units) are available for continuously recording water levels at various key points within the wetland. Constituent Concentrations Table 22.2 summarizes an example of monitoring program in support of operation of a municipal wastewater-polishing wetland treatment system. This list includes measurements of the water quality of all major inows and outows associated with the treatment wetland. Inows include the sources of pretreated wastewater entering the wetland as well as natural inow streams that may have a signicant effect on the water quality or the hydrologic budget of the natural wetland treat- ment system. As described in Table 22.2, the parameter list to be tested at all major inows and outows at least monthly includes all regulated pollutants and integrative measures such as BOD 5 , TSS, pH, dissolved oxygen, water temperature, con- ductivity, NO 2 -N + NO 3 -N, ammonia, nitrogen, TKN, total phosphorus, chloride, and sulfate. The frequency of opera- tional monitoring for system control is dictated by the size and capacity of the system, the sophistication of the owner’s staff and sampling equipment, and site-specic factors related to inuent quality variability and climatic factors. Inow and outow stations may also be monitored less frequently for selected heavy metals or organics that might be present in the wastewater and for whole efuent acute and chronic toxicity. If water quality characteristics are highly variable for any of the inow or outow locations, or if there are weekly or monthly permit limits, sampling should be more frequent than monthly or quarterly. These water quality data as well as water-budget param- eters should be organized and recorded in computerized spreadsheets for visual analysis of variability and trends. The constituent loadings may be compared to design criteria to ascertain if the system is being operated according to design. The constituent mass balances are easily constructed from system data on such a spreadsheet and may be used to deter- mine if design conditions are being achieved, and whether the system is performing as expected. Table 22.2 reects the example of municipal wastewater treatment, but constructed wetland systems are used for a wide FIGURE 22.6 A staff gage provides valuable information on the water depth and volume in the wetland. Photo from the Lake Nebagamon, Wisconsin, treatment wetland system, in the exit deep zone. The screen box covers the intake for the efuent siphon. TABLE 22.2 Example of Monitoring Requirements for Operation of a Large FWS Wetland Treatment System Monitored Parameters Sample Locations Sampling Frequency Flow Inows and outows Daily Rainfall, pan evaporation, and air temperature Adjacent to wetland Daily Water stage Within wetland Daily In the Field Water temperature, dissolved oxygen, pH, and electrical conductivity Inows and outows Weekly Plant cover (dominant species) Near inow, near outow Annually In the Laboratory BOD 5 , TSS, Cl , SO 4 2 NO 2 +NO 3 N, NH 4 N, TKN, TP, and pathogens Inows and outows Monthly Metals, organics, and toxicity Inows and outows Quarterly Note: Some of these are needed for permit reasons, others for mass balancing. © 2009 by Taylor & Francis Group, LLC 774 Treatment Wetlands variety of other purposes, including animal waste treatment, groundwater remediation, and urban and agricultural storm- water. Each of these applications generates a similar require- ment for a strategy to promote understanding. All are based upon knowledge of the water budget, but the suite of chemicals of interest may change, depending upon the application. BIOLOGICAL MONITORING Biological monitoring within a wetland treatment system provides information, concerning the structural integrity (health) of the vegetation and fauna. Protection of this bio- logical integrity is important from an environmental habi- tat perspective and because of the biota’s control of wetland operational performance. It has been suggested that every component of the wetland ecosystem should be monitored, including sediments, water, vegetation, benthic invertebrates, sh, amphibians, reptiles, birds, and mammals (Ecological Services Group, 1996; Wren et al., 1997). This concept origi- nates from the fact that some of these ecological components have been routinely found to be above pristine background for some contaminants, as expected for a treatment system. However, there have been no ecological disasters associ- ated with the thousands of treatment wetlands over the thirty- plus years of the technology, and thus the question involves (possibly) impaired wetland ecology, not acute problems for wildlife. Therefore, biological studies fall in two categories: monitoring needed to foster the design water quality func- tions, and research needed to more fully dene impacts of less-than-perfect water on various organisms that may inhabit the treatment wetland. There is usually an important distinc- tion between constructed wetlands and natural wetlands as recipients of wastewater: the former are usually not subject to intensive biological monitoring but the latter may be in some states (e.g., Florida). A compromise is typically put in place with some facet of biological monitoring or research being conducted in addi- tion to the requisite compliance and mass balance monitor- ing. Typical categories are bird, mammal, and sh studies. Broad-scale potential for ecological problems is commonly assessed by examining or forecasting the concentrations of specic chemicals or toxins. For instance, great care is used in the arid regions of the United States because of the known acute effects of selenium that has impacted wildlife wetlands (e.g., Kesterson, California). Wetlands designed for ancillary benets can have non- treatment goals and requirements. Trends in use of the treat- ment wetlands by plant and wildlife species may indicate needed changes in system management to support those non- treatment goals. Surveys for rare or threatened species may also be conducted when appropriate. MONITORING VEGETATION Cover estimates and observations concerning plant health should be a routine part of operational monitoring in a con- structed wetland treatment system. Because plants grow slowly and are important for maintaining the performance of wetland treatment systems used for water quality treatment, problems must be anticipated and prevented before they are serious or have progressed too far. Reestablishing a healthy plant com- munity in a constructed wetland is a slow process when the plants have been harmed because of operator neglect. Plant growth is monitored by estimating percent cover and average plant height. These nondestructive techniques are used to ascertain the status of plant development before and during wetland treatment system operation. Plant cover is an estimate of the percentage of the total ground area cov- ered by stems and leaves. This parameter can be estimated by walking through or next to a plant stand and visually deter- mining a cover category for the plants. Typically, seven cover categories are sufcient for this visual estimation method: (1) less than 1%, (2) 1–5%, (3) 6–10%, (4) 11–25%, (5) 26–50%, (6) 51–75%, and (7) 76–100%. Cover estimates can be made on a ner scale by using a frame to delineate specic areas. Cover estimates should be made at enough locations in the constructed wetland to pro- vide reasonable statistical averages for comparison between cells and between dates. Ground-level photography may be used to provide archi- val information about vegetation. Photos at identied specic locations give a visual time series of potential succession in the system. For instance, a rodent eatout may be documented ( s ee Figure 19.6 for example). Aerial photography is an excellent supporting tool that can provide estimates of the type and percent cover of veg- etative communities in treatment wetlands. Figure 22.7 illustrates the vegetative patterns for one of the constructed treatment wetlands at the Des Plaines River site near Wad- sworth, Illinois. The various areas of different vegetation are clearly visible, but ground-truthing is required to establish which plants are growing in the zones. 22.3 WATER LEVEL AND FLOW MANAGEMENT There are not many controls on the operation of a treatment wetland, especially Type A systems that mimic passive, natu- ral wetlands. Water depth and water ow rate are the two principal controllable features. FLOW MANAGEMENT For many continuous ow FWS treatment wetlands, there is not a choice of the amount of water that reaches the wetland on a daily basis. However, there are choices for some storm- water systems and for lagoon-wetland combinations. For larger continuous ow systems, there are multiple ow paths and hence decisions on how much water to allocate to these parallel paths. Any system may incorporate recycle, which will be under the control of the operator. Flow Balancing In the case of multiple parallel paths, the operator has the choice of how much of the incoming ow to apportion to each ow path. Often, the adjustment of the proportion is © 2009 by Taylor & Francis Group, LLC Management, Operations, and Maintenance 775 not physically easy but rather involves opening and closing of multiple valves or structures. The performance of each ow path depends upon the amount of water it receives, with the optimum occurring when the hydraulic loadings to the various ow paths are equal, so that no one ow path is underworked or overworked. It is informative to examine the sensitivity of the over all system performance to the ow split or potential imbalanced loading. To do so, consider a hypo- thetical example of a system with two ow paths of equal area in parallel, both of which perform according to the P-k-C* model. For the conditions indicated in Figure 22.8, if the equal fractions of the ow are sent to each of the two paths, the inlet concentration is reduced by about 80%, from 100 to 19.3 mg/L. As more water is diverted to path 2, its performance starts to decline, until the extreme when all the water goes to path 2, doubling its hydraulic loading compared to design. At that extreme, the outlet concentration is about double the design value at 39.3 mg/L (about 60% reduction). Of course, path 1 becomes underloaded, and its performance increases. The combined outow from the system shows that imbalance as well. But interestingly, a modest amount of imbalance is not particularly harmful. For the hypothetical example, a 60–40 split is forecasted to raise the outlet concentration to 20.1 mg/L, about 5% higher than design. The extreme of shutting down one ow path may be necessitated by a need for unusual maintenance activities. For instance, ow paths in STA1W of the Everglades pro- tection project have suffered hurricane damage, requiring shut down of a parallel path for vegetation reestablishment. To the extent that such events may be anticipated, allowance can be made in design. Water may be diverted elsewhere for FIGURE 22.7 (A color version of this gure follows page 550) Wetland EW3 at the Des Plaines River site near Wadsworth, Illinois. The fringe zone is vegetated by cattails and bulrushes, whereas the interior has submerged aquatic vegetation and oating-leaved plants. Open water shows as the dark areas on this false color infra-red photo. 0 5 10 15 20 25 30 35 40 45 0.00 0.10 0.20 0.30 0.40 0.50 Fraction to Path 1 Concentration (mg/L) Path 1 Path 2 Combined Outlet FIGURE 22.8 A hypothetical example of ow imbalance between two parallel paths in a treatment wetland, for the conditions shown in the corresponding table. Path 1 is overloaded at the expense of Path 2. Total Flow 5,000 m /d 100 mg/L Path1Area 3 i QC 100,000 m * 0 mg/L Path 2 Area 100,000 m 2 2 2 C k 00 m/yr Wetland Area 200,000 m 3 TIS 2 P © 2009 by Taylor & Francis Group, LLC 776 Treatment Wetlands treatment if such alternatives exist, or extra area and paths may be included in anticipation of path shutdowns. Systems with Storage Flow management is an important part of those systems that involve seasonal or event storage because the stored water may be discharged according to any of several schedules (see Chapter 17). In all of these systems, the overriding goal is to bleed down the stored water in a timely fashion to make room for the next event or for the next storage season. In small-scale systems with winter storage, cost consider- ations will usually lead to storage that does not much exceed the volume of inow over the cold season. In spring, the oper- ator is faced with selecting the program of withdrawals from storage to be sent to the treatment wetland. Such withdraw- als may be limited by pump capacities, which would also be constrained by cost considerations, or if gravity discharge is possible, the withdrawal rates may be much less constrained. For instance, the Houghton Lake, Michigan, system has a wetland discharge season extending from May 1 to October 15 (170 days), but the transfer pump is capable of transferring the water over about 100 days. Therefore, the operator has discretion over which days there is inow to the wetland, just so long as the storage is at its minimum level at the end of the wetland-operating season. Strategies ranging from paced uniform discharge to fast emptying followed by minimum pond level maintenance have been used (see Chapter 17). The discharge of municipal wastewater treatment lagoons is in some circumstances constrained by the ability of receiv- ing waters to maintain acceptable water quality. Such cases occur when the volume of the discharge is signicant com- pared to a receiving stream ow. In that situation, the stream may become dominated by the efuent, especially at times of low recipient ow. It may then be desirable to vary the dis- charge of treated efuent to avoid unacceptable water qual- ity in the combined ow downstream of the discharge. The ability to vary discharges is in turn conditioned on the ability to store water during periods of low acceptable discharge. The ows in the recipient dictate the magnitude of allowable discharge, and hence this strategy is termed hydrograph con- trolled release (HCR; Kadlec and Pries, 2004). Stormwater systems may be preceded by ow-equalization reservoirs to prevent the low degree of treatment that accom- panies large event ows. Examples of this are currently under design and construction in South Florida, where vast (5–10,000 ha) reservoirs will collect stormwater runoff before passing it through the comparably vast stormwater treatment areas (con- structed wetlands), many of which are already built. These are likely to be HCR systems, serving the water-quantity needs of the downstream-receiving systems, which are the Florida Ever- glades. Operating protocols are currently under development. By p assing It is often not feasible, for hydraulic reasons, to treat large event ows in stormwater wetlands. If the hydraulics allow, treating all the water would generally result in the greater load removal, but will pass higher concentrations to receiving waters for the large events. Hydraulic integrity often requires that extreme ows be avoided or else damage to the wetland may result. Therefore, the design of such systems typically allows for bypass. However, this is usually accomplished by the design of control structures rather than by any overt oper- ator action. Exceptions are the STAs of South Florida, most of which have operator-controlled bypass structures. De p th Management It has been shown that depth is not a crucial factor in the areal removal rates of many wastewater parameters (see Chapters 9 and 10, for example). The extra detention time of deeper water does not foster a higher removal, and in many cases, shal- lower water shows better removals. It is much more important to manage water depth for the health and vigor of the plant community, which is typically sensitive to water depth. DEPTH CONTROLLABILITY In very large FWS wetlands, the hydraulic resistance of the vegetation creates a dilemma for operations. The water depth is not controllable by changes in outlet structure settings, except for a small region near the outlet (see Chapters 2 and 18). Over most of the system, the water ow rate and the veg- etation resistance will control the water depth. Therefore, it is important to distinguish between wetlands that can have depth control via outlet structures and those that cannot. In the latter case, depth control is unavoidably linked to ow control and vegetation management, and not to any modica- tion of structure settings. As an illustration, consider the two hypothetical FWS wetlands for which different level scenarios are calculated in Figures 22.9 and 22.10. Both have the same hydraulic-loading rate of 5 cm/d, and both are presumed to have the same (high) vegetation resistance to ow. First consider a modest size 1-ha wetland with L:W = 4. If operation is at an exit weir setting of 0.5 m, then there is essentially a zero gradient in the water surface (level pool). If the exit weir is then lowered to 0.15 m, the water level goes down everywhere, with only a few cen- timeters of head loss forecast (Figure 22.9; 18.9 cm down to 15.0 cm). The average depth is lowered to 17.1 cm. Next con- sider the long, narrow wetland with L:W = 25. When the outlet weir elevation is 0.5 m, the inlet elevation is 3.9 cm higher and the mean depth is 0.52 m. For this system, when the outlet elevation is lowered to 0.15 m, there is not a drawdown of the entire wetland (Figure 22.10). The inlet depth drops only to 0.48 m, and the mean depth is reduced only to 36.5 cm. For the specied ow, the vegetation is holding back the water and lowering the outlet level cannot change that. Design should be such that water depth control is at the out- let structure or else the operator has no option for depth control except to lower the ow rate. For HSSF wetlands, the hydraulic prole will be governed by both Darcy’s Law and the outlet weir setting. Again, outlet control is the preferred method of regulating the water level (see Figure 21.11). This will require a conservative bed design with a high hydraulic conductivity. © 2009 by Taylor & Francis Group, LLC [...]... there will likely 790 Treatment Wetlands Clogged bed response: Drop at outlet control structure produces no noticeable change in ponding at the influent end FIGURE 22. 21 Clogged-bed water level response in HSSF wetlands (From 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... water-loving types that are of concern in treatment wetlands For HSSF wetlands, minor problems are often incurred by small burrowing rodents; in VF systems, such activities can cause a disruption of the filtering sand layer, leading to a loss of treatment efficiency (Weedon, 2002) There are design techniques that can help in controlling muskrats, nutria, and beavers in FWS treatment wetlands, FIGURE 22. 13... column phosphorus 788 Treatment Wetlands FIGURE 22. 18 This berm at the Hayward, California, treatment wetland is need of maintenance because of erosion 22. 6 MAINTENANCE OF STRUCTURES It is obvious that wetland operation involves keeping the physical features in good condition, including inflow and outflow structures, pumps, piping, berms, and levees Two features of FWS treatment wetlands may be singled... and water level in the wetland For large FWS treatment wetlands, the situation is more complex, because of the size of the structures Automatic selfcleaning trash racks can be used to span inlets that are several meters in width (Figure 22. 20) Removal of the debris may require truck transport 22. 7 LONG-TERM PROSPECTS Part of the natural function of treatment wetlands is the buildup of sediments and soils... can proceed under water This can pose a problem in treatment wetlands, particularly those with deep zones Beavers may construct travel channels to move from one deep zone to another, thus constructing short-circuits (Figure 22. 15) (a) (b) FIGURE 22. 15 Beaver travel channel at the Tres Rios Hayfield treatment wetland near Phoenix, Arizona This short-circuit became evident during this water drawdown... Operations, and Maintenance 785 TABLE 22. 4 Mosquito Density by Site and System in the NADB v.2.0 (1998) Site Name Average Density (#/m3) System System Name Constructed Wetlands Arcata, California Santa Rosa, California Hemet/San Jacinto, California Sacramento, California 1 1 1 1 Arcata Wetlands Kelly Farm Hemet/San Jacinto Sacramento Wetlands 217 915 1,337 1,012 Natural Treatment Wetlands Reedy Creek, Florida... (Walton et al., 1990) Generally speaking, adequately designed and properly operated FWS treatment systems will produce the same amount of mosquitoes as equivalent natural wetlands Imma- © 2009 by Taylor & Francis Group, LLC ture mosquito populations in treatment wetlands typically indicate low average densities (Table 22. 4) Mosquitoes can travel 1–5 km from their hatching location (Service, 1993) Mosquito... paths, connecting open water areas such as deep zones These channels are hydraulic short-circuits, which impair the effectiveness of treatment (Figure 22. 14) Muskrat invasions of treatment wetlands are sometimes quite impressive in terms of numbers Knowlton et al (2002) note that muskrats have been problematic at the 37-ha Columbia, Missouri, project with removal of over 1,300 muskrats in 2.5 years Destruction... attracted to treatment wetlands, both as a breeding location and as a place of refuge Redwing blackbirds (Agelaius phoeniceus) love to nest in cattails and do so in great numbers in treatment wetlands of temperate North America Yellowheaded blackbirds (Xanthocephalus Management, Operations, and Maintenance 779 FIGURE 22. 11 Waterfowl at the Commerce Township, Michigan, municipal wastewater-polishing wetland... populations (May, 2004) As noted in Chapter 19, the Victoria, Texas, wetlands suffered significantly because of herbivory by the nutria, necessitating hunting and trapping A contract trapper was employed at the Mt Angel treatment wetland in Oregon The treatment wetlands at Halsey, Oregon, also encountered nutria eatouts Treatment wetlands in the southern tier of states, plus the east and west coastal . the principal day-to-day activities required to achieve successful performance of these low-tech treatment wetlands. Other operations and mainte- nance activities in treatment wetlands, such. of veg- etative communities in treatment wetlands. Figure 22. 7 illustrates the vegetative patterns for one of the constructed treatment wetlands at the Des Plaines River site near Wad- sworth,. impacted wildlife wetlands (e.g., Kesterson, California). Wetlands designed for ancillary benets can have non- treatment goals and requirements. Trends in use of the treat- ment wetlands by plant