Waste stabilization ponds (WSPs) and their variants are one the most widely used wastewater treatment systems in the world. However, the scarcity of systematic performance data from full-scale plants has led to challenges associated with their design. The objective of this research was to assess the performance of 388 full-scale WSP systems located in Brazil, Ecuador, Bolivia and the United States through the statistical analysis of available monitoring data. Descriptive statistics were calculated of the influent and effluent concentrations and the removal efficiencies for 5-day biochemical oxygen demand (BOD5), total suspended solids (TSS), ammonia nitrogen (N-Ammonia), and either thermotolerant coliforms (TTC) or Escherichia coli for each WSP system, leading to a broad characterization of actual treatment performance. Compliance with different water quality and system performance goals was also evaluated. The treatment plants were subdivided into seven different categories, according to their units and flowsheet. The median influent concentrations of BOD5 and TSS were 431 mg/L and 397 mg/L and the effluent concentrations varied from technology to technology, but median values were 50 mg/L and 47 mg/L, respectively. The median removal efficiencies were 85% for BOD5 and 75% for TSS. The overall removals of TTC and E. coli were 1.74 and 1.63 log10 units, respectively. Future research is needed to better understand the influence of design, operational and environmental factors on WSP system performance.

INTRODUCTION

Waste stabilization ponds (WSPs) and their variants constitute the simplest form of wastewater treatment, especially for small towns (von Sperling & Chernicharo 2005) because they present favorable characteristics, such as satisfactory performance, low cost, low maintenance, and sustainability (Mara 2008; Muga & Mihelcic 2008). The use of WSPs for sewage treatment in the United States dates back to 1901 in San Antonio, Texas (US EPA 2011). Researchers began publishing papers on ponds in 1950, and from this date the development of these systems began growing in the United States, Australia, New Zealand, Israel, South Africa, India, Canada and Latin America, in Brazil, Mexico, Colombia, Peru, Costa Rica, Cuba and Ecuador (Jordão & Pessôa 2014). In Europe, this type of treatment system is also widely used in France (rural areas with less than 1,000 inhabitants), Portugal, Spain, Greece, Italy and Germany (Mara & Pearson 1987).

There are different types of WSPs, and the most common are: anaerobic, facultative, maturation, and aerated ponds. A common arrangement includes two or three ponds operating in series, such as an anaerobic pond followed by a facultative pond; depending on the required quality of the effluent, one or more maturation ponds may follow the facultative pond (Mara 2008). Another common configuration is a facultative pond or an aerated pond followed by one or more maturation ponds. The US EPA (2011) also suggests the use of WSPs in combination with other treatment processes.

The scarcity of consolidated performance data from full-scale WSP systems has led to challenges associated with their design. In most cases, important operational factors that influence the performance are not considered, resulting in systems that may be oversized and too expensive or undersized and incapable of meeting performance standards and complying with water quality and performance goals. The limited availability of effective and consistent performance data for a large set of full-scale systems has created a knowledge gap with regard to the influence of WSP system configuration on the efficiency of treatment. Thus, the creation and maintenance of a database with information of existing systems is considered very important. In this sense, it is expected that this study may assist researchers and practitioners in the interpretation of actual performance of full-scale pond systems with different configurations, regions, climatic conditions and discharge standards.

The objective of this research was to evaluate the performance of full-scale WSP systems, subdivided into different treatment configurations, located in the North and South American continents (Ecuador, Bolivia, Brazil, and the United States), using descriptive statistical analysis of monitoring data for 5-day biochemical oxygen demand (BOD5), total suspended solids (TSS), ammonia nitrogen (N-Ammonia), thermotolerant coliforms (TTC) and Escherichia coli. Compliance with different water quality goals (final effluent concentrations) and targets for system performance (percent removal) was evaluated for each pollutant and for all configurations.

METHODS

A total of approximately 57,700 data about the performance of 388 WSP systems in Brazil (Federal District, Minas Gerais, São Paulo and Rio Grande do Norte); Bolivia (Yungas region); Ecuador (Cuenca) and the United States (Missouri and Georgia) were analysed. These regions were selected due to their differences in climate and the systems were selected because they included WSPs for at least one unit process and had data available from at least five different sampling events over the course of one or more years. The data were obtained from wastewater or sanitation service provider companies, state environmental agencies and research databases. In the data sets obtained from the various sources there was no clear explanation on the sampling and analytical methods adopted at each plant, and it was assumed that procedures followed the Standard Methods or other accepted protocols. When using secondary data, such as in this study, there is always the chance that incorrect values are used. Those which showed obvious deviations from expected reasonable ranges have been excluded, but the authors tried to preserve the database as close as possible to the original ones, without much interference on them and possible exclusions of data due to indirect inferences from expected behavior.

The data were grouped according to seven different WSP system configurations (Table 1). Concentrations of BOD5, TSS, N-Ammonia, TTC, and E. coli were available at the influent and final effluent of each system, with the exception of WSP systems in the United States, where N-Ammonia, TTC and E. coli concentrations were only reported in the final effluent.

Table 1

WSP system configurations established for this research and the corresponding number of systems for each configuration and location

Configuration Acronym Location
 
Total 
Bolivia Ecuador USA
 
Brazil
 
BO EC GA MO MG RN SP FD 
Facultative pond FP   22  64  89 
Anaerobic pond + facultative pond ANP + FPa    38  45 
Facultative pond + maturation pond FP + MP  156  170 
Aerated pond + facultative/maturation pond AEP + FP/MP  13     17 
Anaerobic reactor + facultative pond AR + FP       11 
Anaerobic reactor + other treatment + maturation pond AR + OTb + MP    11  17 
Pond + post treatment P + PTc    39     39 
  Total 1 2 7 234 27 9 107 1 388 
Configuration Acronym Location
 
Total 
Bolivia Ecuador USA
 
Brazil
 
BO EC GA MO MG RN SP FD 
Facultative pond FP   22  64  89 
Anaerobic pond + facultative pond ANP + FPa    38  45 
Facultative pond + maturation pond FP + MP  156  170 
Aerated pond + facultative/maturation pond AEP + FP/MP  13     17 
Anaerobic reactor + facultative pond AR + FP       11 
Anaerobic reactor + other treatment + maturation pond AR + OTb + MP    11  17 
Pond + post treatment P + PTc    39     39 
  Total 1 2 7 234 27 9 107 1 388 

BO, Bolivia; EC, Ecuador; GA, Georgia (USA); MO, Missouri (USA); MG, Minas Gerais (Brazil); RN, Rio Grande do Norte (Brazil); SP, São Paulo (Brazil); FD, Federal District (Brazil).

aSystems in Missouri and Georgia in this category have anaerobic or aerated ponds (alternating between aerobic and anaerobic conditions to improve the removal of nitrogen), followed by facultative lagoons and maturation ponds.

bOther treatment refers to anaerobic filters or facultative ponds.

cRefers to a configuration composed of ponds (anaerobic, facultative, aerated) followed by some other type of post treatment unit process (wetlands, disinfection, etc.).

The following descriptive statistics of the influent and effluent concentrations and the removal efficiencies of the studied parameters (BOD5, TSS, N-Ammonia, TTC and E. coli) for each WSP system were analysed for each treatment plant: arithmetic and geometric means, medians, maximum and minimum values, percentiles of 5%, 25%, 75% and 95% and standard deviations. Because in most systems data availability was only for the influent (raw sewage) and final effluent, without monitoring at intermediate points of the treatment chain, the statistics presented are for each system, as a whole, and not for the individual units (reactors and ponds) that comprise each system. Similarly, removal efficiencies are for the system, and not for each unit in the series.

Using the raw data of each of the WSPs (within each configuration), the descriptive statistics of the concentrations (mean, median, geometric mean, maximum, minimum, standard deviation, percentiles 5% and 95%, etc.) of each parameter were calculated. The medians (for BOD5, TSS and N-Ammonia) and geometric mean (for TTC and E. coli) of each system were used in the descriptive statistics of each configuration (removal efficiencies and concentrations). This was done in order to give the same weight to each treatment plant, regardless of their number of data. Finally, the value of total removal efficiency of each configuration corresponds to the median of the median removal efficiencies from each system. The Kruskal–Wallis one-way analysis of variance (ANOVA) with post hoc multiple comparison test was used to compare the removal efficiencies between the seven configurations (α = 0.05). The Portal Action software (São Carlos, SP, Brazil) add-on for Microsoft Excel was used for all statistical tests.

In order to assess the capacity of the treatment systems to comply with discharge regulations, typical quality and performance goals were established for the final effluent concentrations and removal efficiencies for BOD5, TSS, N-Ammonia, TTC and E. coli, according to common limits presented in the literature (EC 1991; Oliveira 2006; Noyola et al. 2012; US EPA 2013) and required by law (Bolivia 1992, 33 U.S.C. 1251 et seq. 1992; Tulas 2003; COPAM-MG 2008; CONAMA 2011) in the countries or regions where the systems under study are located. Table 2 shows the quality goals established in this study, expressed in terms of values and also a qualitative interpretation expressed in terms of the least to the most stringent standard. The percentages of compliance to these goals were calculated according to Equations (1) and (2): 
formula
1
 
formula
2
where Nec is the number of measurements with effluent concentrations below the target effluent concentration, Ntec is the total number of measured effluent concentrations, Nre is the number of removal efficiencies above the target removal efficiency, and Ntre is the total number of measured removal efficiencies.
Table 2

Quality and performance goals established for effluent concentrations and removal efficiencies

Effluent concentrations
 
Level of goal BOD5 TSS N-Ammonia TTC E. coli 
mg/L mg/L mg/L MPN/100 mL MPN/100 mL 
Most stringent 20 20 1.00E + 02 1.00E + 02 
More stringent 60 60 10 1.00E + 03 1.00E + 03 
Less stringent 100 100 15 1.00E + 04 1.00E + 04 
Least stringent 140 140 20 1.00E + 05 1.00E + 05 
Removal efficiencies
 
Level of goal BOD5TSS% N-Ammonia% TTC Log10 units E. coli Log10 units 
Most stringent 90 90 80 
More stringent 80 80 70 
Less stringent 70 70 60 
Least stringent 60 60 50 
Effluent concentrations
 
Level of goal BOD5 TSS N-Ammonia TTC E. coli 
mg/L mg/L mg/L MPN/100 mL MPN/100 mL 
Most stringent 20 20 1.00E + 02 1.00E + 02 
More stringent 60 60 10 1.00E + 03 1.00E + 03 
Less stringent 100 100 15 1.00E + 04 1.00E + 04 
Least stringent 140 140 20 1.00E + 05 1.00E + 05 
Removal efficiencies
 
Level of goal BOD5TSS% N-Ammonia% TTC Log10 units E. coli Log10 units 
Most stringent 90 90 80 
More stringent 80 80 70 
Less stringent 70 70 60 
Least stringent 60 60 50 

Following a methodology similar to that used by Oliveira (2006), for the purpose of comparison and examination of compliance with the quality and performance goals, three criteria were established. High stipulates that the system has high compliance, with 90% or more of the samples complying with the established goal; Medium states that the system has medium (intermediate) compliance, with 50% to 90% of samples meeting the goals; Low indicates that the configuration has low compliance, with less than or equal to 50% of sampling occasions satisfying the goals. These criteria were used for both the effluent concentration and the removal efficiencies goals for the five parameters: BOD5, TSS, N-Ammonia, TTC and E. coli. The evaluation was made for each WSP system within each configuration and for each parameter separately.

RESULTS AND DISCUSSION

Concentrations and removal efficiencies

Figures 1 and 2 show the influent and effluent concentrations of BOD5 and TTC for each of the seven configurations and the removal efficiencies of BOD5 and TTC (log10 units). Due to the space limitation, it is not possible to show all graphs from the other constituents. TSS were similar to BOD5 graphs and E. coli was similar to TTC graphs. N-Ammonia had insufficient data to lead to conclusive graphs. Median values of concentrations and removal efficiencies are presented in Table 3. Detailed descriptive statistics for each parameter and configuration are provided as annexes in the Supplementary Material (available with the online version of this paper).
Table 3

Median values of influent and effluent concentrations and removal efficiencies for each configuration and parameter

Influent and effluent concentrations (influent/effluent)
 
    BOD5 TSS N-Ammonia TTCa E. colia 
Configuration Acronym mg/L mg/L mg/L MPN/100 mL MPN/100 mL 
Influent concentrations (all configurations) – 431/50 397/47 42/6 2.84E + 07/3.58E + 03 4.94E + 07/1.56E + 04 
Facultative pond FP 492/93 400/150 42/13 2.68E + 07/1.27E + 05 2.66E + 07/6.98E + 05 
Anaerobic pond + facultative pond ANP + FP 453/69 397/122 –/17 3.14E + 07/1.80E + 05 9.19E + 07/2.28E + 05 
Facultative pond + maturation pond FP + MP 410/19 180/32 –/1 –/1.54E + 02 –/6.41E + 01 
Aerated pond + facultative/maturation pond AEP + FP/MP 102/17 132/27 –/2 –/2.94E + 02 – 
Anaerobic reactor + facultative pond AR + FP 439/62 439/62 – 2.72E + 07/– –/2.55E + 05 
Anaerobic reactor + other treatment + maturation pond AR + OT + MP 422/50 – –/11 –/1.21E + 04 –/2.93E + 05 
Pond + post treatment P + PT –/14 –/26 –/1 –/1.67E + 02 –/1.85E + 01 
Removal efficiencies
 
Configuration Acronym BOD5TSS % N-Ammonia % TTCa Log10 units E. colia Log10 units 
Removal efficiency (all configurations) – 85 75 57 1.74 1.63 
(a)Facultative pond FP 78 (a < b); (a < f) 53 (a < b); (a < d); (a < e) 57 1.43 (a < b); (a < f) 1.33 (a < b) 
(b)Anaerobic pond + facultative pond ANP + FP 85 (b > a) 71 (b > a) – 1.90 (b > a) 1.99 (b > a) 
(c)Facultative pond + maturation pond FP + MP 75 39 (c < e) – – – 
(d)Aerated pond + facultative/maturation pond AEP + FP/MP 86 79 (d > a) – – – 
(e)Anaerobic reactor + facultative pond AR + FP 85 86 (e > a; e > c) – – – 
(f)Anaerobic reactor + other treatment + maturation pond AR + OT + MP 86 (f > a) 78 – 1.95 (f > a) – 
Pond + post treatment P + PT – – – – – 
Influent and effluent concentrations (influent/effluent)
 
    BOD5 TSS N-Ammonia TTCa E. colia 
Configuration Acronym mg/L mg/L mg/L MPN/100 mL MPN/100 mL 
Influent concentrations (all configurations) – 431/50 397/47 42/6 2.84E + 07/3.58E + 03 4.94E + 07/1.56E + 04 
Facultative pond FP 492/93 400/150 42/13 2.68E + 07/1.27E + 05 2.66E + 07/6.98E + 05 
Anaerobic pond + facultative pond ANP + FP 453/69 397/122 –/17 3.14E + 07/1.80E + 05 9.19E + 07/2.28E + 05 
Facultative pond + maturation pond FP + MP 410/19 180/32 –/1 –/1.54E + 02 –/6.41E + 01 
Aerated pond + facultative/maturation pond AEP + FP/MP 102/17 132/27 –/2 –/2.94E + 02 – 
Anaerobic reactor + facultative pond AR + FP 439/62 439/62 – 2.72E + 07/– –/2.55E + 05 
Anaerobic reactor + other treatment + maturation pond AR + OT + MP 422/50 – –/11 –/1.21E + 04 –/2.93E + 05 
Pond + post treatment P + PT –/14 –/26 –/1 –/1.67E + 02 –/1.85E + 01 
Removal efficiencies
 
Configuration Acronym BOD5TSS % N-Ammonia % TTCa Log10 units E. colia Log10 units 
Removal efficiency (all configurations) – 85 75 57 1.74 1.63 
(a)Facultative pond FP 78 (a < b); (a < f) 53 (a < b); (a < d); (a < e) 57 1.43 (a < b); (a < f) 1.33 (a < b) 
(b)Anaerobic pond + facultative pond ANP + FP 85 (b > a) 71 (b > a) – 1.90 (b > a) 1.99 (b > a) 
(c)Facultative pond + maturation pond FP + MP 75 39 (c < e) – – – 
(d)Aerated pond + facultative/maturation pond AEP + FP/MP 86 79 (d > a) – – – 
(e)Anaerobic reactor + facultative pond AR + FP 85 86 (e > a; e > c) – – – 
(f)Anaerobic reactor + other treatment + maturation pond AR + OT + MP 86 (f > a) 78 – 1.95 (f > a) – 
Pond + post treatment P + PT – – – – – 

‘–’, no data available.

aColiform values refer to the geometric mean.

Data in parentheses tell us if there is a significant difference or not between the configurations (α = 0.05).

Figure 1

Box-plot of influent and effluent concentrations of BOD5 and TTC of seven configurations.

Figure 1

Box-plot of influent and effluent concentrations of BOD5 and TTC of seven configurations.

Figure 2

Box-plot of percentage removal efficiencies of BOD5 and log units removal for TTC for the seven configurations.

Figure 2

Box-plot of percentage removal efficiencies of BOD5 and log units removal for TTC for the seven configurations.

BOD5

All configurations, except for AEP + FP/MP, showed similar BOD5influent concentrations, with median values that ranged from 410 mg/L to 492 mg/L. In the case of the systems with the AEP + FP/MP configuration, the median concentration was 102 mg/L, which is much lower than values reported in the literature (Faleschini et al. 2012; von Sperling 2014). Although the ponds in this latter configuration may not represent characteristics that are comparable to the other ones, they have been maintained in this study in order to show that this situation also occurs in practice. The influent concentrations obtained for all configurations ranged widely from 92 mg/L to 985 mg/L and the overall median was 431 mg/L. BOD5 values of 400 mg/L would be classified as strong domestic sewage, according to Metcalf & Eddy (2014). For all configurations, the BOD5effluent concentrations were less than half of the influent values. There was a greater variability of concentrations between configurations. The median effluent concentration for all systems was 50 mg/L. The range obtained was from 3 mg/L to 280 mg/L. The removal efficiency ranged from 53% to 96%, with a median of 85%. The ANP + FP and AR + OT + MP were the configurations with best removal efficiency for BOD5. Comparing these values with results from surveys of other pond systems, the overall median values for the influent and effluent concentrations are here lower than those reported by Oliveira & von Sperling (2009) with ponds in Southeast Brazil. Comparing with average values obtained by Recault et al. (1995) in France, the influent concentrations are higher here, and the effluent concentrations are very similar.

TSS

The ANP + FP and FP configurations showed very similar influent concentrations, with great variability, and median values of 400 mg/L and 397 mg/L, respectively. The FP influent concentration showed to be lower than the value reported by Oliveira & von Sperling (2009). For other configurations, the medians were between 132 mg/L and 439 mg/L, and the AEP + FP/MP configuration had the lowest median influent concentration (132 mg/L). The total range for all configurations was from 57 mg/L to 806 mg/L and the overall median was 397 mg/L, which also classifies the influent as strong domestic sewage (Metcalf & Eddy 2014). This overall median value is higher than the average values reported by Recault et al. (1995) and Noyola et al. (2012), whose values were around 260 mg/L. The range of concentrations was much broader than those reported in the literature, but the median observed was lower than that reported in the literature (Oliveira 2006). For effluent concentrations, a wide variation was noted between configurations. The FP configuration showed the widest range, from 29 mg/L to 222 mg/L (percentiles 25% and 75%), and the median was similar to the ANP + FP configuration, with values of 150 mg/L and 122 mg/L, respectively. The configurations with the lowest effluent concentrations were AEP + FP/MP and P + PT, with respective medians of 27 mg/L and 26 mg/L. The overall range obtained was 3 mg/L to 498 mg/L and the overall median was 47 mg/L. The removal efficiencies had a greater variability among the configurations, presenting even negative values, which are likely to be due to the accumulation of algae and its discharge in the effluent, very common in systems with FP and FP + MP configurations. The overall interquartile range of removal was 21% to 90% and the overall median was 75%. AR + FP was the configuration with best removal efficiency for TSS. In terms of effluent concentrations, the overall median value is lower than the value reported by Oliveira & von Sperling (2009). However, in terms of removal efficiency, the overall value is very similar to those reported by these authors.

N-Ammonia

Data for influent concentrations were not available for the majority of systems, as the Department of Natural Resources in Missouri and Georgia (USA) do not require influent concentrations to be reported, and most operators of the systems in Brazil, Ecuador, and Bolivia do not collect these data. The FP configuration was the only one with the minimum data (five) for the analysis, with a median influent concentration of 42 mg/L (ammonia as N), similar to Recault et al. (1995) and Faleschini et al. (2012). Effluent concentrations were reported for all configurations. There was a large variability among the configurations, with a range from 0.02 mg/L to 51 mg/L and an overall median of 6 mg/L; 46% of the effluent N-Ammonia concentrations from all configurations were below 1.9 mg/L, 62% were below 4.8 mg/L, and 92% were below 17 mg/L, which are the thresholds for the maximum average concentrations recommended by the US EPA (2013) for 30-day, 4-day, and 1-h periods, respectively, to avoid toxicity to aquatic life in surface waters. These results demonstrate that some configurations of WSP systems are capable of meeting stringent numeric nutrient limits. The median removal efficiency for the FP configuration was 57%.

TTC and E. coli

Similar to N-Ammonia, there was a lack of information available for the influent concentrations for TTC and E. coli. Based on the data available for the three configurations with complete information, the geometric mean values of the TTC influent concentrations were 2.7 × 107 MPN/100 mL (FP and AR + OT + MP) and 3.1 × 107 MPN/100 mL (ANP + FP), and the total range obtained was from 7.6 × 106 MPN/100 mL to 2.5 × 108 MPN/100 mL, which are similar to ranges reported previously in the literature for Brazil and the United States (von Sperling & Chernicharo 2005; Metcalf & Eddy 2014). The overall geometric mean was 2.84 × 107 MPN/100 mL, which is lower than the values reported in Oliveira & von Sperling (2009) and Noyola et al. (2012). The FP + MP, AEP + FP/MP and P + PT configurations had similar effluent concentrations, with geometric mean values below 300 MPN/100 mL. For the other configurations, the variability in the effluent concentrations was greater. The overall range for all configurations was between 7.1 MPN/100 mL and 4.2 × 106 MPN/100 mL, and the overall geometric mean was 3.6 × 103 MPN/100 mL. The range of removal efficiencies for TTC was 0.9 to 4.6 log10 units, with an overall geometric mean of 1.74 log10 units.

The results for E. coli were very similar to TTC. Only the FP and ANP + FP configurations presented influent concentration data. The geometric mean values of the influent concentrations were 1.06 × 106 MPN/100 mL and 3.11 × 1010 MPN/100 mL, and the overall geometric mean was 4.9 × 107 MPN/100 mL, which are within the range of reported values in the literature (von Sperling & Chernicharo 2005; Metcalf & Eddy 2014). For effluent concentrations, there was a great variability among the configurations, resulting in a very wide range of geometric mean values, from 1.0 MPN/100 mL to 6.2 × 108 MPN/100 mL. The overall geometric mean obtained was 1.6 × 104 MPN/100 mL. The removal efficiency was very similar to TTC, with a range of 1.02 and 3.23 log10 units and an overall geometric mean removal of 1.64 log10 units.

The values of the removal efficiencies for TTC and E. coli indicate that the effluent of systems with the FP and ANP + FP configurations would be suitable for restricted irrigation (of crops not typically consumed raw), using subsurface irrigation; the effluent of the systems with the ANP + FP configuration would be suitable for unrestricted irrigation of high-growing crops (that may or may not be consumed raw) if drip irrigation is used (WHO 2006).

For the P + PT configuration, the effluent concentrations for the parameters studied were always lower than those from the other configurations. The interpretation of the performance of this configuration is difficult and limited, due to the variability of post treatment methods employed. However, it is interesting to present their data and information, in order to highlight that different effluent qualities can be achieved, depending on the post treatment process employed.

WSP are natural treatment systems, whose performance depends on operational and environmental factors. Table 4 shows a summary of the environmental factors (ambient temperature and isolation) for each location and operational factors (hydraulic retention time – HRT – only for the configurations that had at least five systems with this information, and surface organic loading rate – only for primary facultative ponds). This table shows the wide diversity of conditions in the pond systems included in this survey. A study of the influence of these factors on treatment performance has been done as part of this overall research (Espinosa 2016), but was not included in this article due to space limitations. An important conclusion of this part of the study is that none of the factors, taken separately, was able to explain the performance of the pond systems, demonstrating that the performance of these systems is a result from a multitude of factors interacting simultaneously.

Table 4

Range (5 and 95 percentiles) and median (in parentheses) values of environmental factors in each location and operational factors in each configuration

Environmental factors
 
  Location
 
 BO EC GA MO MG RN SP FD 
Ambient temperature (°C) NS 16.6–16.6 (16.6) 10.6–14.1 (11.9) 20.8–24.5 (23.2) 25.8–26.5 (26.2) 21.0–24.8 (23.8) 
Insolation (kWh/m2 d) NS 4.3–4.4 (4.4) 4.0–4.2 (4.1) 4.8–5.5 (5.3) 5.8–6.4 (6.3) 5.0–5.3 (5.2) 
Operational factors
 
  Configuration
 
 FP ANP + FP FP + MP AEP + FP/MP AR + FP AR + OT + MP P + PT 
Hydraulic retention time (d) 13–64 (33) 12–40 (23) 11–143 (72) 21–59 (36)a NA 18–131 (87)b 
Surface organic loading rate (kgBOD/ha d) 95–638 (316) NA 36–831 (344) NA NA NA NA 
Environmental factors
 
  Location
 
 BO EC GA MO MG RN SP FD 
Ambient temperature (°C) NS 16.6–16.6 (16.6) 10.6–14.1 (11.9) 20.8–24.5 (23.2) 25.8–26.5 (26.2) 21.0–24.8 (23.8) 
Insolation (kWh/m2 d) NS 4.3–4.4 (4.4) 4.0–4.2 (4.1) 4.8–5.5 (5.3) 5.8–6.4 (6.3) 5.0–5.3 (5.2) 
Operational factors
 
  Configuration
 
 FP ANP + FP FP + MP AEP + FP/MP AR + FP AR + OT + MP P + PT 
Hydraulic retention time (d) 13–64 (33) 12–40 (23) 11–143 (72) 21–59 (36)a NA 18–131 (87)b 
Surface organic loading rate (kgBOD/ha d) 95–638 (316) NA 36–831 (344) NA NA NA NA 

NS, not specified; NA, not applicable; L, less than five data.

HRT values correspond to the complete system (all units). Surface organic loading rate corresponds to the primary facultative pond.

aValues correspond only to the pond (not considering the HRT of the upflow anaerobic sludge blanket (UASB)).

bValues correspond to the first pond.

Compliance with different effluent quality and system performance goals

The percent of systems in compliance with the established goals for water quality (concentrations of pollutants in the final effluent) and system performance (percent removal of pollutants) was evaluated for each of the seven configurations. Each system was rated as having high (≥90%), medium (>50% and <90%), or low (≤50%) compliance with four different goals (least stringent, less stringent, more stringent and most stringent), as shown in Figures 3 and 4, respectively. When considering the water quality goals for the final effluent concentrations (Figure 3), the P + PT configuration had the best performance of all configurations for all pollutants, which is likely due to the additional unit process used for post treatment in these systems. In Figure 4, despite the lack of information for some configurations, it is clear that most WSP systems have difficulty in meeting the established goals for percent removal of the different pollutants.
Figure 3

Percentage of compliance with the effluent concentration goals: most stringent, more stringent, less stringent and least stringent for the seven configurations, according to the criteria: High (≥90% compliance), Medium (>50% and <90% compliance), and Low (≤50% compliance).

Figure 3

Percentage of compliance with the effluent concentration goals: most stringent, more stringent, less stringent and least stringent for the seven configurations, according to the criteria: High (≥90% compliance), Medium (>50% and <90% compliance), and Low (≤50% compliance).

Figure 4

Percentage of compliance with the removal efficiency goals: most stringent, more stringent, less stringent and least stringent for the seven configurations, according to the criteria: High (≥90% compliance), Medium (>50% and <90% compliance), and Low (≤50% compliance).

Figure 4

Percentage of compliance with the removal efficiency goals: most stringent, more stringent, less stringent and least stringent for the seven configurations, according to the criteria: High (≥90% compliance), Medium (>50% and <90% compliance), and Low (≤50% compliance).

For BOD5, most configurations had low compliance (<50%) with the most stringent goals for effluent concentration (≤20 mg/L) and removal efficiency (≥90%), where more than 85% of the WSP systems had less than 50% of instances with removal efficiencies higher than 90%. For the less stringent BOD5 quality goals, the results were more acceptable for most of the WSP systems. The FP + MP, AEP + FP/MP and P + PT configurations had high compliance with 93%, 100% and 100% of all WSP systems, respectively. For example, 90% of the measurements of BOD5 effluent concentration were lower than 100 mg/L for 93% of WSP systems with the FP + MP configuration. For the other configurations, the compliance is also high but only with 36% or less of all WSP systems. For BOD5 removal efficiency goals, the compliance improves from the less stringent goal (70%); in this goal medium compliance prevails in all configurations with more than 50% of their WSPs. High compliance appears with less than 45% of the WSP systems, except configuration AR + FP that presents high compliance with 67% of their WSP systems. For the least stringent goal (60%), all configurations achieved a high level of compliance (≥90%), except for FP and FP + MP, which reached a medium level of compliance (>50% and <90%).

For TSS, the results showed that the two lowest goals for the final effluent concentration (20 mg/L and 60 mg/L) were difficult for WSP systems to meet; perhaps due to the presence of algae in the final effluent, which is common in these types of systems. Systems from all seven configurations had mostly low compliance for the most stringent level (20 mg/L); systems with the FP + MP, AEP + FP/MP and P + PT configurations achieved mostly high compliance for the more stringent level (60 mg/L), and for the other two goals (100 and 140 mg/L), the compliance was more acceptable for systems with the AR + FP configuration. Systems with FP and AP + FP did not achieve high compliance with even the least stringent goal for effluent concentration. The AR + FP configuration was the only configuration with data that had mostly medium or high compliance with the TSS removal efficiency goals.

For N-Ammonia, there was great variability between the different configurations in terms of their compliance with effluent concentration goals. All configurations had low compliance for the most stringent goal of 5 mg/L, and mostly medium or high compliance for the least stringent goal (20 mg/L); the configurations reached high compliance with more than 40% of their WSP systems, except ANP + FP, which reached high compliance only with 13% of their WSP systems. Analysis of compliance to the removal efficiency goals for N-Ammonia was not possible due to the lack of available influent concentration data for this parameter (Figure 4).

The performance of the configurations to the compliance of the TTC effluent concentration goals was variable. The FP and ANP + FP configurations had the lowest levels of compliance with the established goals. This makes sense, considering that these configurations are not designed to eliminate fecal indicators and pathogens, since they lack post treatment unit processes like maturations ponds, wetlands or disinfection. The FP + MP and P + PT configurations had the best compliance with TTC effluent concentration goals (Figure 3). Other configurations had mostly low compliance (≤50%) with the effluent quality goals of 1.00E + 03 MPN/100 mL and 1.00E + 04 MPN/100 mL. For the least stringent goal (1.00E + 05 MPN/100 mL), the compliance was generally high, with more than 60% of WSP systems from all configurations, except the FP configuration, which still presented mostly low or medium compliance. Furthermore, for the TTC removal efficiency goals (Figure 4), there were only three configurations (FP, ANP + FP and AR + OT + MP) with complete data. The compliance with the two most stringent goals (4 and 5 log10 units) was very low. For the FP configuration, low compliance prevailed in both cases with 100% of WSP systems; systems with the ANP + FP configuration reached medium compliance with the least stringent goal (2 log10 units) for 55% of WSP systems; and low compliance for all other goals with 100% of their WSPs.

Finally, for E. coli, like for TTC, the performance of the different configurations was variable. For the most stringent effluent concentration goal, only the FP + MP and P + PT configurations reached high compliance with 60% and 80% of systems. All systems with other configurations showed low compliance (≤50%) with this most stringent goal. Once again, this demonstrates the need to have a maturation pond or another kind of post treatment unit process in WSPs to comply with goals for the concentration of fecal indicator microorganisms. For the least stringent goal (1.00E + 05 MPN/100 mL), the compliance was better, and the FP configuration was the only one that did not have any systems that reached high levels of compliance. For the removal efficiency goals, it was not possible to evaluate differences between systems with different configurations due to the lack of data available for this parameter at the influent points. Only the FP and ANP + FP configurations had systems with complete information. Systems with the FP configuration had low compliance for all four removal goals, and systems with the ANP + FP configuration had low compliance to the three most stringent removal goals (3, 4 and 5 log10 units), but had 50% of systems reach medium compliance for the least stringent goal (2 log10 units).

CONCLUSIONS

To the authors' knowledge, this is the broadest evaluation to date of the performance of WSP systems, based on the number of plants and treatment configurations. A database with seven configurations of WSPs from 388 different systems was created and analysed. The configuration with the greatest number of systems was facultative pond followed by maturation pond, whereas systems with anaerobic reactors were the least common in the database. The overall median BOD5 and TSS influent concentrations were consistent with strong domestic sewage, and the range of influent concentrations for TTC and E. coli were also consistent with values previously reported in the literature. The effluent concentrations of N-Ammonia, TTC and E. coli demonstrate that some configurations of WSP systems are indeed capable of meeting stringent effluent limits for these parameters. The overall median removal efficiencies for all configurations were 85% for BOD5, 75% for TSS and 57% for N-Ammonia (N-Ammonia results based on the FP alone), and the overall geometric mean removals of TTC and E. coli for all configurations were 1.74 and 1.63 log10 units, respectively. The FP and FP + MP configurations generally had lower removal efficiencies than other configurations, particularly the AEP + FP/MP, AR + FP and AR + OT + MP configurations. Compliance with the most stringent effluent quality and removal efficiency goals was very low, but improved for the less stringent goals. Systems with facultative ponds followed by maturation ponds, and systems with WSPs followed by post treatment unit processes had better compliance with the effluent quality goals. Future research is needed to better understand the design, operational, or environmental factors (such as hydraulic retention time, temperature, and solar insolation) that enable some WSP systems to provide more efficient removal of pollutants than others. This is especially important for the sustainability of wastewater management in tens of thousands of small towns throughout the world that currently rely on this technology (and many more that will rely on it in the future) for wastewater treatment.

ACKNOWLEDGEMENTS

The corresponding author would like to thank COPASA, ETAPA, André Calado, Sílvia Oliveira and Benjamin Drabek, who provided the data for this research. CAPES/CNPq provided funding to support the corresponding author (MFE) and FAPEMIG provided complementary support. This material is also based upon work supported by the US National Science Foundation Graduate Research Fellowship Program under Grant No. 1144244. Any opinions, findings, and conclusions or recommendations expressed in this report are those of the authors and do not necessarily reflect the views of the National Science Foundation.

REFERENCES

REFERENCES
33 U.S.C. 1251 et seq
1992
Federal Water Pollution Control Act
. .
Bolivia
1992
Ley Del Medio Ambiente No. 1333
.
Bolivia
(in Spanish)
.
CONAMA (Conselho Nacional de Meio Ambiente)
2011
Resolução CONAMA N° 430
. .
COPAM-MG
2008
Deliberação Normativa Conjunta COPAM/CERH-MG N.o 1, de 05 de Maio de 2008
(in Portuguese)
.
Espinosa
M. F
.
2016
Avaliação Do Desempenho E Da Influência Das Condições Ambientais E Operacionais de Lagoas de Estabilização, Com Base Em Um Banco de Dados Internacionais
.
Dissertação de Mestrado-Programa de Pós-Graduação em Saneamento
,
Meio Ambiente e Recursos Hídricos da Universidade Federal de Minas Gerais
.
(in Portuguese).
Jordão
E. P.
Pessôa
C. A
.
2014
Tratamento de Esgotos Domésticos
,
7th edn
.
ABES
,
Rio de Janeiro, Brazil
(in Portuguese).
Mara
D. D.
Pearson
H. W.
1987
Waste Stabilization Ponds: Design Manual for Mediterranean Europe
.
Lagoon Technology International, Leeds
,
UK
, pp.
1
53
.
Metcalf & Eddy
2014
Wastewater Engineering: Treatment and Resource Recovery
,
5th edn
.
AECOM, Edinburgh
,
UK
.
Muga
H. E.
Mihelcic
J. R.
2008
Sustainability of wastewater treatment technologies
.
Journal of Environmental Management
88
(
3
),
437
447
.
Noyola
A.
Padilla-Rivera
A.
Morgan-Sagastume
J. M.
Güereca
L. P.
Hernández-Padilla
F.
2012
Typology of municipal wastewater treatment technologies in Latin America
.
Clean – Soil, Air, Water
40
(
9
),
926
932
.
Oliveira
S. M. A. de C.
2006
Análise de Desempenho E Confiabilidade de Estações de Tratamento de Esgotos
.
Tese de Doutorado – Programa de Pós-Graduação em Saneamento, Meio Ambiente e Recursos Hídricos da Universidade Federal de Minas Gerais
,
Belo Horizonte
,
Brazil
(in Portuguese).
Oliveira
S. C.
von Sperling
M
.
2009
Evaluation of the overall performance and relationship with operational variables at 73 full-scale primary facultative ponds
. In:
8th IWA Specialist Group Conference on Waste Stabilization Ponds
.
IWA – International Water Association
,
Belo Horizonte, Brazil
.
Recault
Y.
Boutin
C.
Seguin
A.
1995
Waste stabilization ponds in France: a report on fifteen years experience
.
Water Science Technology
31
(
12
),
91
101
.
TULAS
2003
Texto Unificado de Legislación Ambiental Secundaria Del Ministerio de Ambiente (TULAS)
.
Legal Rule or Regulation
. In:
Norma de Calidad Ambiental Y de Descarfa de Efluentes Al Recurso Agua
(
Ministerio del Ambiente del Ecuador
,
ed.).
Presidencia de la República de Ecuador
,
Quito, Ecuador
(in Spanish)
.
US EPA Environmental Protection Agency
2011
Principles of Design and Operations of Wastewater Treatment Pond Systems for Plant Operators, Engineers, and Managers
.
EPA/600/R-11/088. Environmental Protection Agency, Washington, DC, USA
.
US EPA Environmental Protection Agency
2013
.
EPA 822-R-13-001
.
Office of Water
,
Washington, DC, USA
.
http://water.epa.gov/scitech/swguidance/standards/criteria/aqlife/ammonia/upload/Aquatic-Life-Ambient-Water-Quality-Criteria-for-Ammonia-Freshwater-2013-Fact-Sheet-April.pdf
.
von Sperling
M.
2014
Características Das Águas Residuárias
. In:
Introdução à Qualidade das Águas e ao Tratamento de Esgotos
,
UFMG
,
4th edn.
Universidade Federal de Minas Gerais, Belo Horizonte
,
Brazil
,
1:51–131 (in Portuguese)
.
von Sperling
M.
Chernicharo
C.
2005
Biological Wastewater Treatment in Warm Climate Regions
.
Vol. 1
.
IWA Publishing
,
London, UK
.
World Health Organization
2006
WHO Guidelines for the Safe Use of Wastewater, Excreta and Greywater: Wastewater Use in Agriculture
.
Vol. II
,
WHO
,
Genova, Switzerland
.

Supplementary data