|
A Technical Guide for Wine Producers
|
|
RECENT ARTICLES | WYNBOER HOME
Microbial dynamics in constructed wetlands used for treating distillery wastewater
Keith du Plessis1, Gideon Wolfaardt2 & Lydia-Marie Joubert3
Key words: Biofilm, COD, constructed wetlands, distillery wastewater, microbial distribution, terminal restriction fragment length polymorphism (T-RFLP).
INTRODUCTION
Distillery wastewater has a complex character with a chemical oxygen demand (COD) that may exceed 100 000 mg/L (Yeoh, 1997). These effluents are among the most difficult to treat, due to the presence of sugars, lignin, hemicellulose, dextrins, resins and organic acids (Sangave and Pandit, 2004). In South Africa, water quality legislation and control of effluent discharge necessitates efficient treatment of distillery wastewater (National Water Act No. 36 of 1998).
Constructed wetlands are a relatively unexplored method of wastewater treatment and would appear to be applicable to the processing of distillery wastewater. Wetlands have a diversity of aerobic and anaerobic zones in the water column, at the soil-water interface, and in the root zone of wetland plants (macrophytes) (Reddy et al., 2002). Gradients between aerobic and anaerobic zones provide a diverse range of habitats that support complex microbial communities. These include bacteria, fungi, protozoa and viruses. The role of microbial activity in the functioning of wetlands is widely recognized, and includes degradation of recalcitrant compounds and nutrient cycling. These processes are important determinants of ultimate water quality. The aerobic populations are primarily found in the water column, in the outer regions of the plant detritus layer and periphyton mats in the water column, and on the soil/gravel particle surfaces. The ability of macrophytes to transport oxygen to the rhizosphere enables aerobic populations to persist in this otherwise anaerobic zone. Nevertheless, anaerobic populations are dominant in the subsurface region of the wetland (Reddy et al., 2002), where they mainly exist as biofilms (complex layers of bacteria surrounded by slime matrices) on surfaces and in gravel or soil interstices.
Culture-independent techniques to elucidate differences within both attached and planktonic microbial communities became available following the development of various molecular tools. The terminal restriction fragment length polymorphism (T-RFLP) technique has proven to be a particularly reproducible and accurate tool for community fingerprinting (Liu et al., 1997; Moeseneder et al., 1999). This method can reveal subtle genetic differences between strains and provide insight into the structure and function of microbial communities (Marsh, 1999).
The aim of this study was to trace microbial community structure and function through chemical, visual and molecular techniques in experimental constructed wetlands fed with distillery wastewater.
MATERIALS AND METHODS
Wetland construction and sampling
In 1999 a study was initiated to monitor the potential use of wetlands for the treatment of distillery wastewater. Five wetlands were constructed at a distillery near Worcester in the Breede River valley of the Western Cape. These became fully operational in August 2003 and include two 45 m long x 6 m wide wetlands (one with gravel and one with a soil substrate) both with a 14 day retention time (RT), and three 6 m x 3 m gravel wetlands (4.5, 9 and 18 days RT, respectively). All the wetlands were planted with Phragmytes and Typha species. The effluent was diverted from the final dam in a series of four holding dams that collect discharge from the distillery. At the time of sampling, this water had a COD load =12 000 mg/L. The RT’s of the experimental wetlands were taken into account when sampling wetland outflow to correspond with the respective influent dates (e.g. the outlet sample of the wetland with a 9 day RT was drawn 9 days after the inlet sample was taken). Major chemical properties of the wastewater over the first two months of operation are listed in Table 1. The COD was determined weekly over a period of 12 months by the Reflux Titrimetric method (American Public Health Association, 1998).
Table 1 (Click image to enlarge)
Microbial community diversity in the experimental wetlands
DNA was extracted from the planktonic (unattached/free-floating) microbial communities, as well as from biofilms on the surface of substrate particles in various zones of the five constructed wetlands. The 16S rRNA genes were amplified via the polymerase chain reaction (PCR) as described by Du Plessis (2006) using the primer pair: 6-carboxyfluorescein (FAM)-labelled 341f (Muyzer et al., 1993) and unlabeled rPP2 (Rawlings, 1995) for Eubacteria. The PCR products were digested with the restriction enzyme Alu1. Digested 16S rDNA was purified and mixed with an internal size standard whereafter the fluorescently labelled terminal restriction fragments (T-RF) were detected on an automated gene sequencer.
RESULTS AND DISCUSSION
Wastewater treatment efficiency
After 1 month, COD removal in the 45 m long soil and gravel wetlands was ~50% and 20 - 40%, respectively (data not shown). After 12 months, however, the average COD removal in the gravel wetland was 84% while COD removal efficiency in the soil wetland decreased to <40%, probably due to excessive plant growth, and clogging of the soil matrix as a result of biomass accumulation, EPS (exopolymeric substances or slime) production and/or insoluble biogas formation by soil microbes (Mattison et al., 2002). COD removal efficiency, averaged over 12 months, was positively correlated with RT in the 6 m long gravel wetlands. The 6 m long gravel wetlands with the 18 day RT was the most efficient (79%), followed by those with 9 (64%) and 4.5 days (39%) (data not shown). Although the wetlands were effective in reducing the concentrations of most of the elements (e.g. K and N) they were not effective in reducing electrical conductivity as assessed over the 12 month period. The N and K were taken up by the wetland plants and stored in above-ground plant tissues i.e. shoots, leaves and stems (Zingelwa, 2007).
Wastewater at the inlet of the constructed wetlands in August and September 2003 was classified as very high- and medium-sodium water, respectively (United States Department of Agriculture, 1954). Very high sodium water is unsatisfactory for irrigation purposes while medium-sodium water will be hazardous in soils with a high cation exchange capacity (CEC), especially under low-leaching conditions. In September 2003 the sodium adsorption ratio (SAR) of the wastewaters at the outlets of all the wetlands, with the exception of the 45 m long soil wetland, was above 5 and therefore not suitable for irrigation purposes. The wastewater at all but one of the sampling points fell into the medium-salinity water category. The exception was the wastewater at the outlet of the 45 m long soil wetland, which was classified as high-salinity. Medium-salinity water can be used for irrigation purposes if a small amount of leaching occurs. High-salinity water on the other hand must not be used on soils with restricted drainage (United States Department of Agriculture, 1954).
Microbial community structure in experimental wetlands
The T-RFLP analyses of the 5 experimental wetlands revealed notable variation in their microbial communities. There was little similarity between the community fingerprints for the respective sampling points over time (data not shown). There was, for example, only 6% correspondence (similarity) between fragments in the wastewater at the outlet of the 45 m long soil wetland between two consecutive months (August and September) while 39% of the fragments in the wastewater at the inlet corresponded over this period. The highest correlation between samples collected on different dates was for those obtained from roots in the wetlands; e.g. 60% similarity over time between biofilm communities on the roots at the outlet of the 45 m long soil wetlands. A possible explanation for this variation, especially in the aqueous phase, is the change in distillery effluent composition discharged into the wetlands (Table 1) which may selectively stimulate and suppress growth of different microbial populations. Overall, these results are in agreement with the study by Ragusa et al. (2004), in which indicators of biofilm development were monitored in constructed wetland microcosms, and it was found that micro-scale biofilm properties changed continuously.
There was little agreement between the correlating T-RFs obtained from the planktonic communities of the in- and out-flowing effluents from the five experimental wetlands in September 2003 (Fig. 1A and B). The highest percentage of correlation that was observed, compared to the in-flowing planktonic community, was that of the soil biofilm community near the inlet of the 45 m long soil wetland (43%; Fig. 1B). No correlation between RT and community profiles was observed when the effluents of the five experimental wetlands were compared. These results imply that various biotic and abiotic factors influence microbial community composition in wetlands, and that unique micro-environments are established in response to the factors (e.g. the optimum RT, pH etc.) that facilitate microbial degradative processes.
Planktonic communities differed markedly from attached communities within the same region of the wetland - e.g. there was only 17% and 18% correlation between the T-RFs obtained from the wastewater at the outlet, and in biofilm communities growing on the gravel and roots, respectively, at the 45 m long gravel wetland outlet (Fig. 1A). Also, planktonic microbial communities in the wastewater at the outlet of the 45 m long soil wetland showed only 26% and 22% similarity with biofilms from soil and roots in the corresponding zones of the wetland (Fig. 1B). The attached communities on different surfaces within the same region of the wetland also varied markedly from each other, with the 45 m gravel wetland showing the highest similarity (45%) between microbial communities on plant roots and gravel support near the outflow (Fig. 1A). The overall highest level of correspondence (87%) occurred between attached communities on the soil in the inflow and outflow zones of the soil wetland (Fig. 1B). The microstructure of the soil substrate, where pores are readily clogged by EPS development, thereby restricting water flow, may lessen the shock effect of effluent composition and thereby stabilize the attached populations.
Vacca et al. (2005) found that the microbial communities in wetland systems are strongly influenced by the movement of the wastewater through the systems, by the substrate used in the wetlands, as well as by the plants. Their study also revealed that fingerprinting of the rhizosphere of plants grown on sand, or on expanded clay, exhibited many differences. This showed that different microbial communities may exist depending on the type of soil in the systems. Chiarini et al. (1998) showed that the rhizosphere of maize plants is markedly affected by the type and parameters of the soil. Similar results were found in this study with the molecular fingerprint patterns of the 45 m long gravel wetland being completely different from that of the soil wetland, showing the responsiveness of natural microbial communities. This adaptability suggests that microbial communities maintain metabolic function by modifying species composition in response to fluctuations in their environment.
Retention time may play a role in the make-up of the microbial communities as the three 6 m long wetlands all had different microbial fingerprint patterns at the outlets (data not shown). There were only 21%, 26% and 8% similarity between T-RFs representing the planktonic microbial communities present in the in-flowing wastewater and out-flowing wastewater of the three 6 m long wetlands (RTs of 4.5, 9 and 18 days, respectively). Mayo and Mutamba (2004) found that improved nitrogen removal occurred with increase in hydraulic RT in a high-rate algal pond, as well as in a subsurface-flow gravel bed constructed wetland. The microbial nitrification-denitrification processes are responsible for the bulk of nitrogen removal (Sikora et al., 1995). Therefore, a change in retention time can alter the microbial community present within a system, which in turn impacts on the effluent amelioration efficiency.
CONCLUSIONS
This study confirmed that gravel is a much more suitable substrate for constructed wetlands than soil, the soil wetland becoming inefficient after 12 months of operation. This loss of efficiency was probably due to excessive plant growth, clogging of the soil matrix due to biomass accumulation, slime production and/or insoluble biogas formation by soil microbes. The soil and gravel wetlands promoted the development of completely different microbial community fingerprint patterns at given sampling dates indicating that the substrate used in constructed wetlands plays a major role in determining the microbial composition. This study showed that a highly dynamic microbial community exists within wetlands. Wetlands can efficiently remove COD even though a low degree of similarity exists between microbial communities in various zones of the same wetland, and even though low similarity exists between communities sampled from the same zone over time. It was striking that planktonic communities differed markedly from attached communities within the same region of the wetland. It would therefore, be difficult to determine wetland system health by means of microbial analysis alone. Future studies should be directed at identifying and monitoring indicator organisms that could be used in combination with chemical and visual analyses in an attempt to determine wetland system health.
ACKNOWLEDGEMENTS
The authors are grateful to Winetech and ARC Infruitec-Nietvoorbij for financial support. Franscos Baron and the rest of the staff at the Soil Science division at ARC Infruitec-Nietvoorbij are acknowledged for their technical support.
For further information contact Keith du Plessis at (021) 809-3158 or e-mail dplessisk@arc.agric.za.
REFERENCES
American Public Health Association, 1998. Standard methods for the examination of water and wastewater. 20th ed., Washington DC.
ABSTRACT
Microbial distribution and activity were studied in experimental systems constructed to evaluate the potential use of wetlands for the treatment of distillery wastewater. Two 45 m, land three 6 m long wetlands, varying in retention time and support medium (soil and gravel), were monitored at a distillery near Worcester in the Breede River valley of the Western Cape. Of specific interest were the dynamics of attached and planktonic microbial communities at various points in the wetlands, as was chemical oxygen demand (COD) removal efficiency. Whole-community DNA samples showed a high diversity among microbial communities that were sampled at 14 locations in the 45 m long wetlands. COD removal was more efficient in gravel than in soil wetlands. In the gravel wetland COD removal was positively correlated with retention time.
|
|
Wynboer is incorporated in WineLand, magazine of the SA wine producers.
Visit our sister sites:
|
| UP |
COPYRIGHT (C) 2000 WineLand |