Wetlands for water pollution control have been in use for more than 100 years. Both Kadlec and Wallace [1
] and Vymazal [2
] give detailed accounts of the major developments in the field over the past century and describe the field as evolving, over time, towards more intensely engineered designs and process train layouts for water treatment optimization purposes. The term “constructed wetland” (CW) is generally used when a wetland for water pollution control has been explicitly built and designed for a water treatment purpose. The term “treatment wetland” (TW) generally encompasses constructed wetlands but is more expansive and includes natural wetlands that are intentionally used as catchment areas for water pollution. Kadlec and Wallace [1
] describe several early 1900s North American examples of natural wetlands receiving water pollution, by design, over extended periods of time as a general polishing (tertiary) water treatment step.
] classifies TW systems based on vegetation (emergent, submerged, floating leaved, and free-floating) and hydrology (free water surface, and subsurface flow), with subsurface systems classified into horizontal and vertical flow, and hybrid systems involving any combination thereof. Free water surface systems are generally more common in North America and Australia [3
], where subsurface flow systems are more common in Europe and were recently described as one of the most common extensive water treatment processes in the world [4
]. Although many design variations exist, often three main (more commonly implemented) configurations are discussed and compared in the literature being free water surface flow (SF), horizontal subsurface flow (HF), and subsurface vertical flow (VF). SF systems are more often used for tertiary water treatment applications, where both HF and VF are more commonly used for secondary water treatment (although many exceptions do exist). VF systems are also sometimes used in primary treatment applications (termed “French systems”). Engineering augmentation, design, and intensification is currently an active area of research with aerated systems being implemented for full-scale applications in many locations worldwide (see Nivala et al. [5
The study of the use and utility of plants in water treatment systems is generally credited to German scientists in the 1950s, the first being Seidel [6
] (as described in [1
]). More recently, the rhizosphere (subsurface zone of interaction between root structures and microbial communities) has been described as a “sunlight driven hotspot” for the degradation of organics [7
], and has been shown to degrade, remove, immobilize, and/or transform a range of contaminants including but not limited to nitrogen compounds, organics, pharmaceuticals, petrochemicals, chlorinated solvents, pesticides, explosives, heavy metals, and radionuclides to name a few [8
]. Removal mechanisms involved in water treatment in TWs include a number of physiochemical, plant related, and microbiologically mediated mechanisms, the host of which are outlined in Kadlec and Wallace [1
]. TWs are relatively straightforward in their implementation. However, their internal operation is quite complex. Microbial communities can be found throughout TWs, however, three main areas are commonly identified: attached, within close proximately to, or associated with roots (rhizospheric); within biofilms surrounding the general media; or in the free water (for SF) or interstitial water (for HF or VF). As wastewater passes through a TW the chemical constituents can be considered food for microorganisms. As this food is utilized by microbial communities in specific areas of the TW, they can anchor and create fixed biofilms through secretion of extracellular polymeric substances [8
]. As this anabolic action of biofilm and microbial mass creation continues, pore space in HF or VF systems can change (on a local-scale) thus driving water through a slightly different hydrological regime, and therefore overall flow-path in the TW. This action is quite interesting as the nutrient flux to biofilms in areas of reduced pore space is thus reduced, naturally limiting biofilm growth. As this process continues over time a subsurface TW can develop to a point where biofilm is either well distributed or perhaps heterogeneous. Stratification of biomass in subsurface flow systems is well documented [10
]. At some point an equilibrium between attachment and detachment (due to local-scale velocity and shear stress) can occur allowing for steady state operation with no bio-clogging [11
]. This steady-state operation is however sometimes not possible due either to simple solids build up, or in part due to the nutrient loading which can drive the microbial community towards anabolic based processes (i.e., creation of biomass/biofilm) which can eventually lead to clogging of pore spaces (see Nivala et al. [12
]). Recent modelling exercises have included these dynamics into fundamental TW investigations [11
], and predict either heterogeneous or homogenous clogging depending on the specific model used.
As described, microbial communities are said to facilitate water treatment through metabolic actions resulting in the general degradation of waste constituents. This occurs through using the waste constituent for either cellular mass and reproduction (anabolism), or energy (catabolism). These metabolic actions are facilitated through a different series of enzymatic reactions based on the specific functional requirement/action of the community, or subset of the community. Different areas within TWs house different environments, which help drive and select for certain functional abilities of inherent TW microbial communities [4
]. It is through these diverse and iterative sets of different conditions, in which wastewater passes through, that TWs are thought to have exceptional and sometimes surprising water treatment potential [8
]. For example, CWs have shown promise for the removal of pesticides [17
] and emerging contaminants, including antibiotics, pharmaceuticals and personal care products [18
Weber and Gagnon [8
] describe four classes of microbial community assessment: activity, enumeration, function, and structure. Enumeration methods look to provide quantitative data on the number or amount of microorganisms in a sample. Activity measurements do not directly account for the number of microorganisms in a sample, but rather look to understand how metabolically active they are. This activity is often directed at a specific type of transformation, for example CO2
evolution implying catabolic respiration and the mineralization of organic material. Functional analysis looks to profile the overall function of a microbial community over a range of metabolic transformations. This in some ways can be thought of as gathering an understanding of many transformation-specific activity measures to build a full picture of the microbial community overall function. Function can also be assessed through piecing together gene pathways available (DNA) or active (mRNA) in a microbial community through molecular techniques. Structural analysis is focused on what microorganisms are present in a sample (often at the individual level, or as operational taxonomic units) and is completed through many techniques ranging from light microscope profiling to metagenomics sequencing. Weber and Gagnon [8
] further define structural analyses to include methods where communities are compared based on structural components, even if not all individuals or taxonomic units are expressly identified. Using this premise, community based DNA fingerprinting methodologies are categorized as a type of structural analysis. Truu et al. [10
], Faulwetter et al. [16
], and Weber and Gagnon [8
] provide an account regarding the different microbial community assessment methodologies available and their general utility. However, a solid quantitative survey of what specific microbial community assessment methods are being utilized in the field of TWs, and for what purpose has not been reported.
The objective of this study is to provide a comprehensive and precise overview of exactly what microbial community assessment methodologies are being used and developed for use in the field of TWs, and for what purposes. To this end a meta-analysis was completed in order to quantitatively and concisely summarize these aspects throughout the history of TWs. This study is not meant to comment on or recap the findings of all microbial community assessment studies completed in the field of TWs, rather to gain an historical perspective on microbial community assessment in TWs, understand what tools are currently available in the field of TWs, and to provide perspective and comment on current and future developments.
2. Literature Review
Literature was gathered and reviewed in two separate phases. The first phase was aimed at understanding past practices and developments with respect to microbial community analysis in the field of TWs. The second phase was aimed at understanding current practices and state of the art in microbial community assessment in TWs.
2.1. Phase One—1880 to 2012
For phase one all potentially relevant peer reviewed journal publications were gathered by searching through the years of 1880 to 2012 using the databases of Compendex (engineering focused), and Web of Science (more generally captures fundamental science journals). Keywords used for this search were: wetland, constructed wetland, treatment wetland, which were then individually combined with the word microbiology, microbiological, or microbial, giving a total of nine searches. The key word “wetland” was designated as useful over this time period as many early publications did not identify systems as “treatment wetlands” or “constructed wetlands” yet they certainly involved the study of a wetland used for water pollution control. These nine individual searches were then combined and duplicates removed leaving a total of 512 papers. Conference proceedings and non-TW related papers were then removed (206 total) which included a large number of studies investigating natural wetlands where no water pollution control was involved. After this, an additional 91 papers were removed, as they did not include the direct study of a microbial community. The latter removal step was found to be a common occurrence as many publications included the word “microbiology”, “microbiological”, or “microbial” within the paper in order to help discuss and contextualize water treatment results, but the authors did not actually directly investigate the inherent TW microbial community using any methodologies in that specific publication. Pathogen removal focused papers also fell into this category as many looked at the removal of microorganisms from influent water, but did not investigate the inherent TW microbial community. Following this process, the remaining 215 publications were then reviewed, with specific content tracked for further analysis. An excel spreadsheet was used to track the publication title, year, focus area of the study, and specific microbial community assessment methodologies used. Methodologies were then classified into the following categories: activity, enumeration, function, and structure (a full listing of methodologies used and their classifications will be presented later). Focus areas were identified based on the type of water contaminant(s) or research area being investigated and included: nitrogen, organics, phosphorus, methane, sulphur, agricultural, chlorinated volatile organic compounds (cVOCs), petrochemicals, BTEX (benzene, toluene, ethylbenzene and xylene), pesticides, emerging contaminants (ECs), metals, and microbial fuel cells (MFCs).
2.2. Phase Two—2013 to 2016
Phase two was conducted in a similar fashion to phase one, but was augmented in several key ways in order to gather additional info to fully understand the current state of the art and general research efforts currently being expended in the field of TWs. Compendex and Web of Science were again used however only across the years of 2013–2016 (which included up until the month of July in 2016), and the searches did not use the keyword “wetland”. It was deemed that the field of TWs was well established by 2013 and in order to gain insight into TW specific state of the art, catching all wetland related studies (i.e., natural wetland studies without water pollution control involved) was not desirable. A total of 1445 publications were gathered. Table 1
summarizes the specific search match results. After removing duplicates, a total of 589 publications remained. In reviewing these 589 publications, 420 were later removed, as they did not utilize a specific microbial community assessment methodology, leaving 169. An interesting note is that phase two gathered a large number of review articles. In addition, a large number of articles were found which reviewed or described alternative water treatment methodologies that then also referred to TWs in some fashion. This was interpreted as evidence that the TW field certainly has matured to a point where it is thought to be a mainstream and useful comparison reference for other novel and developing technologies.
The remaining 169 publications were then reviewed for content and information tallied in an excel spreadsheet. In addition to the same suite of information gathered from publications in phase one, phase two also gathered information on the region where the study was undertaken, the system configuration/design, and the system size. Region was captured as the country from which the corresponding author resided. This was the only way to gather consistent information throughout the entire process, however it is identified that small biases may exist. For example, the corresponding author of Button et al. [19
] is from North America, even though the actual systems studied are in Europe. System configuration was gathered through the authors own system configuration identification, or from provided schematics and other descriptions. In some cases, informed estimations were required. For example, a large number of publications which did not identify the system configuration surrounded the investigation of large-scale systems (often more than 10 ha in area), in those instances the author chose to identify the systems as surface flow, as it was deemed unlikely that systems of such size would be subsurface vertical or horizontal flow. The system size was quantified in terms of total system volume in litres (L). The system sizes were then further categorized into four size ranges: microcosm (0.1–10 L), mesocosm (10–1000 L), pilot-scale (1000–100,000 L), and large-scale (>100,000 L). In many cases, system depth was not given. In those instances, the author chose to use a standard assumed depth of 1 m for all system types (SF, HF, or VF). In the majority of cases, only publications focused on pilot-scale or large-scale systems chose to not report a system depth, and therefore the majority of systems where an assumed depth of 1 m was used happened to be surface flow (which admittedly means 1 m may be an overestimation in some cases). There were a small number of cases where the authors chose to not report size dimensions of any sort. In those instances, an implied system size identification was sought, such as the authors describing their system as “pilot-scale”. In order to estimate a size from an implied size description, the author chose a value one order of magnitude above the lowest value for that size category: 1 L for microcosms, 100 L for mesocosms, 10,000 L for pilot-scale, and 1,000,000 L for large-scale.
4. Conclusions and Future Horizons
TW research takes many forms. Where in past decades one could attend the WETPOL (Wetland Pollution Dynamics and Control) and the IWA: Wetlands for Water Pollution Control conferences and generally be up to do date on all research progress occurring, the current amount of research activity in the field of TWs is beyond the point where this is possible. TWs are becoming truly internationalized, and this means the range of applications will most certainly increase with new situations and challenges yet to be discovered. This meta-analysis showed that TW research involving the assessment of microbial communities is on the rise, especially since 2013. SF, HF, and VF system types are all being investigated in approximately equal proportions worldwide.
High powered metagenomics sequencing methods are being applied, microarrays have been used to map functional pathways, and stable isotope analysis has been used to assist in functional analysis. The field of TWs truly is at the edge of microbial community analysis technological implementation. mRNA has also been extracted from lab-scale TW systems. This is quite a step. mRNA can be exceptionally difficult to extract from environmental samples [42
]. If mRNA extraction for TWs is possible on a wider scale, micro-array and qPCR analysis could begin to return active function information. Although exciting, this type of extraction and methodology will require significant development and expertise to be applied on a wider scale.
Detachment of representative microbial community samples is a basic question that has not received sufficient attention in the field of TWs. When assessing biofilm originating from gravel, roots, sediment, or otherwise, one wants to gain a representative sample for analysis, yet not bias the community in the process. This can be challenging. Two example studies are Weber and Legge [52
], which investigated the effects of different detachment methodologies on culturable bacteria from CW gravel, and Button et al. [53
], which evaluated the effects of sample preparation on CLPP specifically. Neither Weber et al. [52
] or Button et al. [53
] were able to find absolutely perfect methodologies, but did provide an evaluation of the suitability of commonly utilized methods/techniques and were able to provide general recommendations. Similar evaluation studies have not been completed for DNA extraction in TWs specifically. Methodological advancements are expected and therefore studies similar to Weber et al. [52
] and Button et al. [53
] are anticipated and recommended for the future. Such studies would help researchers standardized their sampling/extraction methodologies allowing for more appropriate and accurate comparisons between studies.
In relation to challenges surrounding the extraction/detachment of representative microbial community samples (especially from biofilm matrices), a similar challenge exists when deciding when and where to take a microbial community sample from a system. In the majority of investigations, the researcher is interested in gaining samples which are representative of the microbial communities contributing to the water treatment capabilities of a TW, however communities from different regions in a TW could contribute in different ways at different times. Since 2013, 33 spatial studies and 20 temporal studies have been published, marking good progress. However, with the vast array of system modifications, intensifications, and hydrological regimes being now employed, in order to gain a solid understanding of microbial community dynamics in all situations, additional effort is required. Understanding both spatial and temporal microbial community dynamics is central to comparing different studies to each other, and in being confident in the results of any single study.
Studies are increasingly becoming multiphasic, meaning more than one class of microbial community assessment is implemented in the same study (63 total from 2013 to July 2016). Multiphasic analysis is enhancing the impact of research by offering multiple lines of evidence. Structural and metagenomics analysis has developed rapidly and is currently the most popular microbial assessment tool in TWs. It is suggested that structural analysis be a part of multiphasic studies or extended to include interpretation of the functional potential of species identified to better assist in connecting the microbial community observations to TW water treatment.
Rather than recommend any specific microbial community assessment methods the author would like to emphasize the concept of “fit for purpose”. The most important question to ask before choosing a method is, “What information do I require, and how will that information help in meeting the overall objectives of the study”. With the understanding that resources are often limited for any single research endeavour, method selection can become challenging. TWs are one of the most complex water treatment systems in existence, and it is for this reason that truly rigorous studies are so difficult to complete because there are so many interdependent aspects at play. In the context of gaining an understanding of the microbial community, using an exceptionally new and expensive method by default can sometimes be a detriment and limit researchers to investigating only one aspect of the microbial community. Those same resources could often be better spent in performing a multiphasic study looking at perhaps all four aspects in some way (enumeration, activity, function, and structure), rather than a single aspect alone. Converging lines of evidence is something all researchers strive to achieve when publishing findings. Striving for multiphasic experimental designs is perhaps the single most important recommendation to be made here.
The field of TWs is truly multidisciplinary. For example, the study of Lünsmann et al. [41
] required collaboration among departments of Biotechnology, Proteomics, Environmental Biotechnology, Isotope Biogeochemistry, Metabolomics, and a Microbial Interactions and Processes Research Group. It is expected that additional collaborations will be developed in the coming years and that TW microbial community research will be multiphasic reaching beyond the current horizons and starting new paradigms.