Comprehensive evaluation of substrates in vertical-flow constructed wetlands for domestic wastewater treatment

Constructed wetlands (CWs) are engineered treatment systems that have been designed and constructed to utilize the natural process of physical, chemical and biological synergistic action among substrates, plants and microorganisms, but do so within a more controlled environment (Vymazal 2011). Today, CWs are recognized as a reliable wastewater treatment ecological technology with the advantages of good purification effect, low construction cost and operation energy consumption, convenient maintenance and management, as well as remarkable ecological benefit, and they represent a suitable solution for the treatment of various wastewater around the world, such as domestic sewage, industrial wastewater, non-point source agricultural pollution, polluted surface water (Wu 2008; Vymazal 2011).

The substrate occupies almost the whole volume in CWs structure, which is the important distinguish between constructed and natural wetland (Wu 2008). Not only do substrates provide physical and chemical support for wetland plants, surface areas and nutrients for microbial attachment and hydraulic condition for sewage flow (Calheiros et al. 2008; Soric et al. 2011; Morvannou et al. 2013), but they also remove kinds of pollutants directly by the action of filtration, adsorption and precipitation and so on (Ballantine & Tanner 2010; Vohla et al. 2011; Saeed & Sun 2012). So substrate is the important influencing factor of purification capacity and stable operation in CWs. In general, the fee of substrate material accounts for a high proportion in the construction cost of CWs. In a word, the selection of appropriate substrate is an important step when designing and constructing CWs.

Traditionally, reed beds had been constructed with local soil as substrate. However, this had caused problems with overland flow and short-circuiting of the wastewater between inlet and outlet because of the low hydraulic conductivity of soils. Therefore, present adopted design guidelines are based on sand or gravel and in some cases intermittently loaded vertical-flow beds instead of horizontal-flow beds (Brix et al. 2001). However, with a seemingly suitable hydraulic conductivity, sand or gravel were known to be poor in phosphorus removal (Brix et al. 2001; Babatunde & Zhao 2009). Then the selection of the material with a high phosphorus binding capacity, used either alone or mixed with sand or gravel as substrate, plays a crucial role in CWs design (Prochaska & Zouboulis 2006; Wu et al. 2011). It is also possible to establish a separate exchangeable filter containing these substrates to remove phosphorus (Lee et al. 2010). Therefore, numerous researches around the world have focused on the feasibility of some natural materials, by-products and man-made products used as substrate in CWs, such as dolomite (Prochaska & Zouboulis 2006), anthracite (Wu et al. 2011), alum sludge (Babatunde & Zhao 2009), steel slag (Lee et al. 2010; Wu et al. 2011), light expanded clay aggregates (Brix et al. 2001). Saeed & Sun (2012) gave a comprehensive review of such materials for organic and nitrogen removal, and Ballantine & Tanner (2010) and Vohla et al. (2011) gave reviews for phosphorus removal.

So far, most of full-scale CWs still adopted local soil, sand or gravel as substrate, such as in China (Yang et al. 1995; Xie et al. 2012). However, substrate selection is void of complete evaluation system, but only depends on the experience derived from some conventional principles and requirements. For example, Wu (2008) proposed that substrate should meet the requirements of enough mechanical strength and chemical stability, as well as rough angular surface. The materials used as substrate in vertical flow CWs should have good hydraulic characteristics, consistent and sustained high phosphorus removal capacity, and be cheap and locally available. In addition, it should be usually desirable that upon phosphorus saturation, such substrates should be amenable to being used as a slow phosphorus release fertilizer (Babatunde & Zhao 2009).

In this study, various factors influencing substrate selection in CWs were turned over based on analytic hierarchy process (AHP), and evaluation system with hierarchical structure was established including four layers, eleven indicators and nine schemes. Then combined with Delphi method and fuzzy synthetic evaluation approach (FSEA), zeolite, anthracite, shale, vermiculite, ceramic filter material, gravel, steel slag, bio-ceramic and combination substrate (isopyknic layered anthracite, bio-ceramic and zeolite) were evaluated comprehensively from the viewpoints of purification effect, practical performance and economic analysis. The quantitative assessment results would provide a basis for substrate selection of CWs design.

AHP is a multiple criteria decision-making tool proposed by T. L. Satty in the 1970s that has been used in almost all the applications related with decision-making (Vaidya & Kumar 2006). It consists of an overall objective, a series of options or alternatives for reaching the objective and a group of factors or criteria that relate the schemes to the objective. The criteria can be further divided into sub-criteria as many hierarchies as require (Xu et al. 2013). It provides a methodology to calibrate the numeric scale for the measurement of quantitative as well as qualitative performances (Vaidya & Kumar 2006). Wan et al. (2013) adopted AHP to assess pesticide pollution management strategies for use in rice production systems in Shanghai, China.

Delphi method has proven a popular tool in information systems research for identifying and prioritizing issues for managerial decision-making (Okoli & Pawlowski 2004). Suggestions of carefully-selected experts are consulted on the questions set in advance, and then answers are analyzed and concluded to give feedback to the experts. The above process is terminated after reaching a pre-defined criterion (e.g. numbers of rounds, achievement of consistent and stable results, Xu et al. 2013). Sustainability of urban wetland resources were assessed using AHP and Delphi method (Xu et al. 2013).

FSEA is designed to group raw data into several different categories according to predetermined quality criteria, which can be normally described using a set of functions that are designed to reflect the absence of sharp boundaries between each pair of adjacent criteria, which was used in the identification of river water quality (Chang et al. 2001).

The data in this study came from measured data in simulation test; fuzzy data on the basis of experience and expert consultation, as well as market survey data (see details below).

Comprehensive evaluation system was established based on AHP, which comprised four different layers: objective layer, criterion layer, indicator layer and scheme layer (Table 1). The objective of substrate evaluation in CWs followed the criteria of purification effect, practical performance and economic analysis. For domestic wastewater treatment, the indicators of purification effect mainly referred to conventional index of water environmental standard, including environmental quality standards for surface water (GB 3838-2002) and discharge standard of pollutants for municipal wastewater treatment plant (GB 18918-2002). Considering the source location of substrate material, volume cost was the sum of price and freight per unit volume substrate. On the basis of the previous studies (Zhang 2007; Li et al. 2010; Wu et al. 2010a; 2010b), zeolite, anthracite, shale, vermiculite, ceramic filter material, gravel, steel slag, bio-ceramic and combination substrate were chosen as nine schemes.

AHP and Delphi method were combined to determine indicator weights. Experts were required to sort criteria and indicators under the same criterion, and then judgment matrixes of pair-wise layers were constructed according to the sorted results. The questionnaires were delivered to 28 experts, of which 23 experts were answered, and 20 questionnaires were valid. The expert group was constituted of researchers, technicians and managers on CWs. Maximum eigenvalue and corresponding eigenvector of each judgment matrix were calculated using square root method. When the consistency test passed, normalized eigenvector were the weights of indicators relative to the upper layer.

Table 2 showed the indictor weights of pair-wise layers and each indicator over objective layer, and the sum of weights in the same layer was 1. The weight of phosphorus removal was the highest, followed by nitrogen removal, chemical stability, and the weight of mechanical strength was the lowest.

Purification effect was the most important factor among three criteria, and the weight of economic analysis was the lowest. Because the main purpose of constructing CWs in this study was domestic wastewater treatment, and substrate expense was one-time cost without regard to replacing substrate after several years. Substrate may play the greatest role in phosphorus immobilization in CWs (Ballantine & Tanner 2010; Vohla et al. 2011), meanwhile substrate adsorption may be an important pathway for NH₄⁺-N removal (Saeed & Sun 2012). The influence of substrate on organic matter removal was fairly small under medium or long term operation, since substrate surface gradually generated relatively stable biofilm over time (Wu et al. 2010a). The removal of suspended solids and indicator bacteria were generally fairly high and with no obvious difference among kinds of substrates (Zhang 2007; Li et al. 2010). So the order of indicator importance under the criterion of purification effect was phosphorus removal, nitrogen removal, organic matter removal, indicator bacteria removal, suspended solids removal and pH value.

Chemical stability of substrates affects the toxicity and harmfulness of CWs effluent, which lead to secondary pollution. Substrate choice in CWs is of major importance as it serves as the support of the living organisms (Calheiros et al. 2008; Soric et al. 2011). Particle size distribution of substrates has a decisive effect on gap size and water capacity, and then is the important factor influencing the clogging in CWs, whereas material type of substrates has a little effect on the block (Zhang 2007). So indicator importance under the criterion of practical performance was in the order of chemical stability, biocompatibility, anti-clogging and mechanical strength.

The data of the indicators under the criterion of purification effect and the indicator of anti-clogging adopted the measured data from simulation test (Zhang 2007; Li et al. 2010; Wu et al. 2010a; 2010b). Three indicators of mechanical strength, chemical stability and biocompatibility under the criterion of practical performance had no clear boundary and were not easy to quantitate. Hence, combined with experience data and the results of expert consultation, five evaluation levels relative to the three indicators were established based on FSEA (Table 3). The data of the indicator of volume cost used the data of market survey.

Collected data generally had its own dimensions and distribution range, so as not to be compared and calculated directly. The data were applied dimensionless method corresponding to various functions according to the different types of the indicators.

The indicator of pH value was normal in the interval of 6–9, which was the indicator of interval type. The other indicators under the criterion of purification effect were the indicators of benefit type, so were three indicators of mechanical strength, chemical stability and biocompatibility. The indicator of anti-clogging used decline rate of porosity as value was the indicator of cost type, so was the indicator of volume cost.

Nine substrates were sorted in accordance with the scores: combination substrate, anthracite, steel slag, zeolite, bio-ceramic, ceramic filter material, gravel, shale, vermiculite (Table 4). Vertical analysis showed that the removal of phosphorus, indicator bacteria and organic matter by anthracite were relatively high, so were nitrogen removal by zeolite and organic matter removal by bio-ceramic. Meanwhile, zeolite and bio-ceramic adapted to biological growth. So zeolite, anthracite and bio-ceramic were three ideal substrates in CWs. However, the removal of nitrogen and suspended solids by anthracite, phosphorus removal by zeolite, and nitrogen removal by bio-ceramic were relatively low. At the same time, volume cost of zeolite was relatively high. So the three substrates had their own disadvantages. The combination of anthracite, bio-ceramic and zeolite was the optimal choice of nine substrates for complementary advantages and synergistic effect. The removal of phosphorus and indicator bacteria by steel slag were relatively high, while its volume cost was relatively low as by-products of metallurgical industry. Nevertheless, chemical stability of steel slag was not stable. Phosphorus adsorption and precipitate with metal oxide and hydroxide on the surface of steel slag was its mechanism of phosphorus removal, and the reason for strong basicity of CWs effluent (Lee et al. 2010). Elevated pH value would be incompatible with sensitive water bodies used as final effluent receivers discharging, as well as generate problems with the growth of commonly applied plants and attached microorganisms (Prochaska & Zouboulis 2006; Lee et al. 2010). Therefore, steel slag was not suitable to be filled into CWs alone, which could be used either in combination with other substrates or by constructing a separate filter for improving phosphorus removal (Lee et al. 2010; Wu et al. 2011).

Horizontal analysis showed that the pH value of CWs effluent used the other eight materials as substrate, except for steel slag, maintained in standard permission range (GB 3838-2002 and GB 18918-2002). Nitrogen removal by combination substrate, zeolite and ceramic filter material were relatively high, so these three materials could be used when nitrogen was the specific pollutant. Meanwhile steel slag and anthracite could be used for removing the specific pollutant of phosphorus. The selection of long-term high-efficient substrates (e.g., zeolite, anthracite, steel slag) was conducive to the decrease of covering area and the extension of service life. Zeolite and vermiculite was suitable for biological growth, so that could be filled on the surface of CWs. Cheap shale and gravel could be used in the areas with underdeveloped economy.

The comprehensive evaluation system of substrates in CWs could be improved according to actual condition. For example, when the objective of constructing CWs was the treatment of special wastewater including heavy metal, persistent organic pollutants, or pharmaceuticals and personal care products, the removal of these pollutants should be brought into the indicator system. On the other hand, for the treatment of Qiantang river water, nitrogen removal should be given a higher weight and volume cost should be lower, in view of that nitrogen was the main pollutant and Hangzhou was the region with developed economy.

The comprehensive evaluation system of substrates in vertical flow CWs for domestic wastewater treatment was established based on AHP, which included four layers, eleven indicators and nine schemes. Phosphorus removal, nitrogen removal, chemical stability were the main factors to be considered for substrate optimization in vertical-flow CWs for domestic wastewater treatment. On the basis of collected data, and combined with Delphi method and FSEA, Combination substrate was considered as the best substrates. Zeolite, anthracite and steel slag were ideal substrates used in different actual condition. The comprehensive evaluation system of substrates in CWs could be improved through the modification of indicator system and indicator weights according to treatment object, project site, and so on.

This research was financially supported by the National Nature Science Foundation of China (No. 51208498) and Major Science and Technology Program for Water Pollution Control and Treatment of China 12th Five-Year Plan (No. 2012ZX07101-007-005). The authors thank laboratory colleagues for their assistance in this study.