Aquaculture Sludge Removal and Stabilization within Created Wetlands (using Vetiver Grass).

From Vetiver Newsletter #16 (November 1996)

Steven T. Summerfelt1, Paul R. Adler2, D. Michael Glenn2, and Ricarda N. Kretschmann1
The Conservation Fund's Freshwater Institute, P.O. Box 1746, Shepherdstown, West Virginia 25443, USA; 2USDA-ARS, 45 Wiltshire Road, Kearneysville, West Virginia 25430, USA

Introduction Removal of solids or nutrients from the effluents of fish farms is often required because of priority and regulations given to minimizing the effect of the discharge on the environment (Ewart et al., 1995). Aquaculture systems offen have two separate discharges, and solids and/or nutrients in both, if left untreated, can have a negative affect upon receiving waters. When systems have two separate discharges, the effluent of largest volume usually contains comparatively low Ievels of solids and nutrients, particularly nitrogen and phosphorous. A second effluent, generated during clarifier backwash, is comparatively small but contalns high levels of concentrated organic solids. The settleable fraction of solids are often removed with settling basins to produce a siudge that is about 5% solids. Many states in the US classify and regulate aquaculture sludge as an industrial or municipal waste, because the sludge is a residual product of wastewater treatment; however, other states consider the sludge to be an agricultural waste, because it is composed of manure and uneaten feed and is thus considered to be a non-toxic nutrient source (Ewart et al., 1995).

Aquaculture effluents such as these, however, do not have to be considered liabilities, because these effluents can be used as inputs for production of other products and used to improve overall facility sustainability. To better achieve sustalnability in aquaculture, The Conservation Fund's Freshwater Institute has been working with the USDA/ARS Appalachian Fruit Research Station on a project titled, "Aquaculture Linked to Plant Culture: Products and Processes." Research has been focused on developing technology to treat nutrients or biosolids in aquaculture effluents while producing other valuable products such as high-value fruits and vegetables (Adler et al.,1996 a; b; in press a), grass turf (Adler et al., in press b), and organic composts (Adler et al.,1996b). Although reuse of effluent streams is always worth considering, it is sometimes difficult to develop the technologies and markets required to support reuse as a form of effluent treatment.

The two most common methods used to recycle solid wastes from aquaculture facilities are land application and composting (Ewart et al., 1995). According to Ewart et al. (1995), land application of manure and other organic wastes (including wastewater) to fertilize agricultural crops is governed in most states by guidelines or regulations that limit the amount of pathogens, heavy metals, and other contaminants and the land application rates. In particular, application rates are based upon nutrient content, soil type, and plant nutrient uptake characteristics to prevent runoff or groundwater contamination (Chen et al., 1991; Ewart et al., 1995). Odor problems can also limit land application in populated areas. Sludge transport from the facility to another point of disposal or reuse is a major factor in the costs of sludge management, because the thickened sludge is greater than 90% water (Black and Veatch, 1995; Reed et al. 1995).

Depending on an aquaculture facility's location and the local regulations, an aquaculture facility may have only limited and costly options avallable for sludge disposal. If land application is not available adjacent to the facility, on site treatment of the concentrated solids discharge with an uncomplicated, low-maintenance plant-based system could reduce solids disposal costs (Outwater, 1994).

Created horizontal flow wetland (HFW; i.e., overland flow wetland) systems have been used with some success to treat high-strength aquacultural wastewaters (Pardue et al., 1994) and other agricultural, municipal, or industrial wastewaters (reviewed by Reed et al. [1995]). HFW systems are usually operated with a hydroperiod to produce cycles of inundation and dewatering. However, HFW systems typically are not loaded with thickened sludges.

On the other hand, constructed vertical-flow wetland (VFW) systems have been used, over the past 20 years to treat thickened sludge (1-7% solids) produced in the clarifier undefflow at wastewater treatment plants (Hofmann, 1990; Lienard et al., 1990; Nielson, 1990, 1993; Riggle, 1991; Outwater, 1994; Reed et al., 1995). VFW wetlands are generally referred to as "reed beds" because they are often planted with reeds. When used for municipal treatment, these wetlands are loaded with 7-10 cm of 2% solids approximately once every 7-21 days (about 30-60 kg/m2/yr). During operation, a series of vegetated beds receives sequential batch applications of sludge. The sequential batch applications are such that the more recently flooded VEW cells are dewatering, wffile beds with older sludge applications are drying. Intervals between sludge addition allow for dewatering and drying. Plants facilitate dewatering by conducting water along their stem and root paths through previous sludge layers and by removing water through evapotranspiration (Outwater, 1994; Reed et al., 1995). The plants also increase biological stabilization of the solids by transporting oxygen to their root zones. Reed bed treatment system have been reported to have a useful lifetime of up to 10 years (Outwater, 1994; Reed et al., 1995).

Aquaculture sludges are good candidates for use in both crop or created wetland. However, if transportation costs make sludge disposal on crop land uneconomical, disposing of the sludge on-site within created wetlands might be the next best alternative. The objectives of the work reported in this paper were to investigate disposal and treatment within created wetlands of an aquaculture sludge produced during clarifier-back-wash. This research focused on the variables controlling capture and stabilization of solids within created wetland systems. Solids removal and stabilization were investigated within two types of created wetlands where water flowed either: (1) vertically, down through a porous substrate; or (2) horizontally, over soil and through hedges. These two wetland types differed in both physical characteristics and in hydraulic distribution and collection.

Both created wetlands types were planted with vetiver grass (Vetiveria zizanioides). Vetiver grass was selected because it is tolerant of a wide range of environmental conditions, and has been proven to control soil erosion throughout the world (Becker, 1992) when planted as narrow hedges, the dense vetiver shoots act as a filter, allowing water to pass through while holding soil back to settle by gravity, thereby preventing erosion.

Vetiver also has an extensive and deeply growing root system that would help malntains the bed's hydraulic conductivity and contribute to oxygen transport into the bed.

Methods Sludge used in these studies was collected from the recirculating trout-production system at the Freshwater Institute (Heinen et al., 1996). Sludge originated from the clarifier backwash and was collected and thickened to about 5% solids in a septic tank before it was pumped to the greenhouse where the wetland cells were located. However, the manner in which the sludge was collected and pumped from the septic tank to the equalization tank within the greenhouse diluted the sludge to about 0.75% dry solids by weight. Sludge pumped from the equalization tank was thoroughly mixed before it was applied to the wetland cells. Solids loading onto both horizontal and vertical wetland types was about 30 kg/m/yr. About 60 l of sludge was applied 6 times daily per wetland cell, approximately evety day from May 12, 1995, until February 18, 1996. No draln and dry period was provided for either type of wetland. However, on three occasions, flow to the HFW cells had to be discontinued for several days to prevent water levels from over-flowing the vessels. Flow rates to each wetland were checked three times per week. Occasionally, a plugged distribution pipe kept sludge from being applied to a given wetland cell.

Six 3.7 x 1.2 x 0.8 m (L x W x H) wetland cells were used to provide three replicates for both types of vetiver beds.
The VFW cells (Figure 1) are sand drying beds planted with vegetation. The VFW cells consisted of a 10-cm layer of sand and three layers of increasingly larger gravel to support the sand over a flow collection pipe (Figure 1), based on criteria provided by Cooper (1993). Sludge was distributed across the top of each VFW cell through a 2.5-cm inside diameter pipe (Figure 1). Solids were trapped on and within the sand as the flow passes vertically through the bed. A 7.5-cm inside diameter drainage pipe at the bottom of the bed collected and carried the flow from each VFW cell. Each VFW cell sloped 2% down to the point where the draln pipe exited the tank. Vetiver tillers were planted at about 15-cm intervals across the entire top of each VFW cell.

The HFW cells (Figure 2) were designed to have the flow travel overland, passing horizontally along the tank's long axis, from one narrow end of the cell to the other. The HFW cells were loaded to a depth of 51 cm with a local topsoil. Rooted vetiver shoots were planted close agalnst each other in three 35-cm wide rows; each row was oriented perpendicular to the long axis of the vessel, and each row was about 61 cm apart (Figure 2). About the same number of vetiver tillers were planted in a HFW cell as in a VFW cell. Sludge was distributed at the upper end of the tank onto a brick to disperse the energy of the flow. The flow passed through the vetiver hedges in the process of traveling from one end of the wetland to the other (Figure 2). The dense shoots of mature vetiver hedges were expected to enhance solids removal by straining and settling. After passing horizontally through the wetland cell, the flow was collected in a perforated draln pipe placed at the end of the cell's long axis and buried under sand and three supporting layers of gravel. Each HFW cell sloped 2% down to the point where the drain pipe exits the tank.

Data was collected on influent and effluent concentrations of total suspended solids (TSS), total volatile solids (TVS), total and dissolved chemical oxygen demand (COD), nitrate, dissolved phosphate, total nitrogen, and total phosphorus. Data was collected on 11 separate weeks from June through February. TSS and TVS were measured using standard methods (APHA, 1989). Total and dissolved COD were measured using a Hach spectrophotometer test kit (Loveland, Colorado). In water samples, nitrate and phosphate were quantified by ion chromatography (APHA, 1989) as described by Adler et al. (in press b). After chemical digestion, total kjeldahl nitrogen (TKN) and total phosphorus were determined by ion chromatography as described by Adier et al. (in press b).

Sludge depths and sludge samples were also taken from each wetland at the end of the 1-yr study and were analyzed for percent volatile solids.

Results And Discussion Results indicated that sludge removal and stabilization occurred within both wetland types (Tables 1 and 2). The VFW and HFW cells, respectively, removed 98 and 96% TSS, 91 and 72% total COD, and 81 and 30% dissolved COD (Table 2). Because little dissolved COD was expected to be removed by physical mechanisms, the increased removal of dissolved COD within the VFW cells was likely due to better anaerobic digestion occurring within the sand and gravel layers of the VFW cells. Both wetland types removed most, 82-93%, of the dissolved phosphate, total kjeldahl nitrogen, and total phosphorus (Tables 3 and 4). Nitrate was produced in both wetland types; however, there was much more nitrate in the effluent from the VFW cells than from the HFW cells (Table 3). Particulate phosphorus were the major form phosphorus in the treated effluent from both wetland types (Table 3). Nitrate was the major form of nitrogen leaving the VFW cells (Table 3).

Nitrate production (Tables 3 and 4) indicates that there was some aerobic bacterial activity (e.g., nitrification) in both types of wetland cells. Although the saturated regions of both VFW and HFW cells were mostly anaerobic, localized aerobic conditions may have been created within wetlands through either root transport of oxygen or by aeration of the flow as it trickied through the gravel-support layers within the vertical flow wetlands. The lower gravel layers were not saturated with water due to the large void spaces between large pieces of gravel. However, much more nitrate was produced in the VFW cells than in the HEW cells, probably because oxygen was transferred from the atmosphere as the flow trickled through the aerated gravel-support layers. Denitrification probably accounted for the removal of some nitrate from both wetland types, but the low level of nitrate in the effluent from the HFW cells may have been due to both insufficient oxygen transfer for nitrification, and to anaerobic conditions that caused rapid denitrification of nitrate when it was produced.

At the end of the study, depths of accumulated sludge in each wetland averaged 11 and 8.1 cm in the VFW and HFW cells, respectively. Although the density of the accumulated sludge was not measured directly, the sludge that accumulated within the VFW cells was less dense than the sludge that accumulated within the HFW cells due to the presence of large voids (air pockets) within the sludge from the VFW cells. Additionally, these sludges contained an average of 43 and 37% volatile solids, respectively. In comparison, the fraction of volatile solids in the sludge that was treated was about 83% volatile (Table 1), and it was 57 - 65% volatile in the treated wetland effluents (Table 2). Therefore, considerable mineralization occurred in the accumulated sludge.

Resistance to water flow through the wetland cells was greater within the HFW cells than in the VFW cells, as indicated by deeper water ponded above the HFW surface (on average 12-18 cm deep) than above the VFW surface (on average 5-12 cm deep). Additionally, we suspect that most of the water flowed horizontally above the soil and across the HFW cells and then filtered through the sand layer covering the collection pipe at the end of the cell. The distance the sludge had to flow horizontally and the thickness and number of hedges in the HFW cells were probably inadequate to physically remove most of the solids. These conclusions were supported by observations of the vetiver hedges and the sludge distribution across the top of the HFW cells at the end of the experiment, which indicated that the three hedges planted across each HFW cell did not develop stem and root masses thick enough to trap most of the solids. Performance may have been enhanced by allowing hedges to thicken more before application of sludge began. Therefore, we think that the similar and favorable particulate removal found in both the HFW and VFW cells were largely due to the sand layers that cover the effluent collection pipes within each wetland cell.

In this research, solids were loaded onto both horizontal- and vertical-flow wetland cells semi-continuously at a rate of 30 kg/m2/yr. Sludge was loaded on the wetland cells at about the same rate as others have recommended for wetland drying beds (Hofmann, 1990; Lienard et al., 1990; Nielson, 1990, 1993; Riggle, 1991; Outwater, 1994; Reed et al., 1995); however, sludge used in this experiment was relatively dilute (0.75% dry solids) when compared to the thickened sludges (1-7%) these same others reported. Additionally, the semi-continuous application of sludge in this experiment meant that only a small volume of sludge was distributed at any given application. Over a two-week period, the more dilute sludge concentrations applied (i.e., higher water content) resulted in a higher hydraulic loading rate than others generally applied to VFW cells (Outwater, 1994; Reed et al., 1995). After the first few weeks of operation, the hydraulic loading used in this experiment always maintalned a flooded condition. Maintaining surface flooded conditions was our original intent when we selected semi-continuous applications. We expected that, when flooded, the sand layer of the VFW and the soil within the HFW would make effective anaerobic filters, which proved true. This hydraulic loading strategy was contrary to conventional wisdom, as others have recommended altemating flooding and drying intervals to enhance plant growth and sludge stabilization by air- and photo-oxidation (Hofmann, 1990; Lienard et al., 1990; Nielson, 1990, 1993; Riggle, 1991; Outwater, 1994; Reed et al., 1995). It is generally held that an aerobic environment helps to minimize odors, breaks down organic matter more rapidly, and makes phosphorus less susceptible to leaching than would anaerobic conditions. However, it is also generally believed that an anaerobic environment stabilizes sludge to its minimum solids mass and requires less energy (e.g., trickling filter height, blower/aerator power) than an aerobic environment. Additionally, this study showed that the the anaerobic-sand filter proved effective at removing dissolved organic molecules.

At the conclusion of the experiment, root growth was observed when all material was removed from the wetland vessels. Root growth was thick below the vetiver all the way to the base of the 51 cm sand and gravel or soil media. Roots had even grown into the bottom drain pipes and had surrounded the bottom layers of large gravel suffiently to make manual gravel removal much more difficult.

Vegetation played an important role in dewatering the sludge, as evapotranspiration accounted for 12-20% of the water balance across both types of wetland cells during the summer months as others have also reported (Outwater, 1994; Reed et al., 1995). Plant growth was vigorous from spring until fall, when all but the base 20-25 cm of plant stem was cut and removed from all wetland cells. Much of the vegetation senesced through the winter, but shoot growth was occuring in portions of the wetlands by the end of February, 1996, when the experiment was terminated. It was apparant from the pattern of uneven shoot growth that occured in all three VFW cells, however, that some factor had limited plant revegetation within the lower third of each HFW cell. There was a total lack of vegetation in these regions. It is uncertain why revegetation did not occur in the lower regions of the VEW cells. However, because both the sand surface and the vessel base of each wetland cell had been sloped 2% down to the drain, an additional 5 cm of ponded sludge (when flooded) had accumulated at the lower end of each cell. It is possible that the additional sludge, along with the continuous anaerobic digestion, ammonia production, and flooded conditions were critical factors that limited revegetation in the lower regions of the VFW cells. Therefore, in future studies we hope to investigate the impact of hydroperiod on revegetation and solids removal and stabilization within created wetlands.