Semiquantitative microbiological risk assessment for water reuse in agriculture: a case study in Brazil Open Access

Although Brazil has about 12%–16% of the total amount of water available in the world (Avins 2015), this availability is not evenly distributed across the territory (Lima et al. 2021). Urban centers in the central region of Brazil have pressure on water resources classified as high, in addition to extensive land use and land cover for agriculture (SEMAD 2020). Therefore, water reuse from treated wastewater comes as an alternative way to meet the demands for non-potable purposes in the region, such as irrigation.

However, due to the constituents present in wastewater after treatment, there is an associated risk for the health of workers, consumers, local population and the environment (Shoushtarian & Negahban-Azar 2020). To assess the hazard related to microbiological constituents when reusing water, fecal contamination indicators are used, such as bacteria of the coliform group (Rebelo et al. 2020; Zhiteneva et al. 2020).

Water reuse must be managed to minimize and eliminate the risks of exposure to the hazard (WHO 2006). Hazard is understood as chemical, biological or physical agents with the potential to cause damage (Vuppaladadiyam et al. 2019). The risk is obtained by the relationship between the hazard, the receptor vulnerability and the possible damage (Rebelo et al. 2020).

Risk assessment is developed to achieve acceptable levels of risk, with different approaches being possible depending on needs and data availability (WHO 2016). Risk assessment involves the likelihood that human exposure to one or more pathogens will result in an adverse health effect (Seto et al. 2018). In cases of non-potable water reuse, methodologies with a semiquantitative or qualitative approach are more suitable (Zhiteneva et al. 2020), such as the model developed by Rebelo et al. (2020), based on the recommendations of ISO 16075-1:2015. Therefore, the aim of this research is to carry out a semiquantitative microbiological health risk assessment regarding theoretical application of agricultural reuse.

For the hazard, the input data is the treated wastewater Escherichia coli density and the treatment level, given away by the Saneamento de Goias S/A – SANEAGO (water and wastewater utility). The utility performs punctual collection at the exit of the WWTP and bimonthly monitoring of the parameter. The analysis was performed according to the Standard Methods for the Examination of Water and Wastewater (SMEWW), method SMEWW 9223 B. The E. coli density data used to determine the hazard of each city represent the 2019 annual geometric mean, and the raw data is in the Supplementary Material.

The microbiological risk assessment methodology developed by Rebelo et al. (2020), predicts, in the first place, the use of E. coli as input, however, it allows for other parameters to be used, and it is enough to develop a new scale of relationship between the parameter concentrations or densities and the equivalent hazard importance factor, as well as is performed in the Table in Figure 2.

According to Rebelo et al. (2020), the risk is classified as despicable (0<risk<3), acceptable (3≤risk<7) or unacceptable (7≤risk<9). Despicable risk should be understood as low risk. If the risk is unacceptable, it is necessary to adopt mitigation measures, such as hazard reduction by adopting more advanced levels of treatment than the first ones, or damage reduction, with the use of a greater number of barriers or more efficient barriers.

Data on the treatment system in the WWTP of each city and the treated wastewater E. coli density, in Colony Forming Units (CFU) per millilitre (mL), are presented in Table 1.

Based on the E. coli density data (Table 1) the hazard importance factor used was 9 (E. coli >10⁴ CFU/100 mL, as shown in Figure 2).

Brazil does not have legislation on standards for water reuse, but there are national guidelines, Interaguas and the recommendations from the World Health Organization (WHO). INTERÁGUAS (2017) indicates a maximum E. coli density of 10³ CFU/100 mL for restricted agricultural reuse (irrigation of crops not intended for human consumption). However, Santos & Vieira (2020) argue that the proposed limits are too restrictive for the Brazilian reality. For restricted low-tech and labor-intensive agriculture reuse, WHO (2006) recommends an E. coli density of 10⁴ CFU/100 mL, determined by a Quantitative Microbiological Risk Methodology (QMRA). QMRA is an assessment based on local characteristics (WHO 2016). In other words, the WHO use general data that might not portray the conditions in Brazil.

Several countries such as Spain, France and Italy have quality standards for water reuse at the national level (European Commission 2016), while other countries such as the United States of America have legislation at the local level (Angelakis et al. 2018). According to Santos & Vieira (2020), in Portugal, in addition to defining standards for E. coli, the legislation bases the quality requirements on a fit-for-purpose approach, recommended by ISO 16075-1:2015. This approach considers the adoption of multiple barriers, which can effectively promote a microbiological reduction or only represent this reduction through barriers. Therefore, the quality limits established according to the fit-for-purpose approach are adaptable to the local reality, being more consistent than the guidelines proposed by WHO (2006) and INTERÁGUAS (2017). It is important to highlight that reuse patterns in agriculture in general, defined by regulations, vary depending on the type of crop and irrigation method.

For the safe practice of agricultural reuse, the following must be considered: workers and farmers exposed to reclaimed water; food consumers; those involved in the marketing and processing of irrigated products; and the population located close to the irrigation site (Shoushtarian & Negahban-Azar 2020).

The receptors theoretically considered for the risk assessment were (Figure 2):

• Farmers who work in the pasture and sugarcane areas to be irrigated;

• Water reuse transport workers by tank truck;

• Local community that inhabits or circulates in the vicinity of the irrigated area.

For sugarcane irrigation, the following were also analyzed:

• Sugarcane industry workers.

The classic exposure routes for water reuse are ingestion, inhalation and dermal adsorption. The first two assume a factor of absolute importance (9) as they are demonstrably harmful to human health. On the other hand, dermal adsorption assumes a factor of weak importance (3) as there is little evidence of damage to health (Russo et al. 2020).

The same exposure scenarios were defined for the neighborhood of the four WWTP analyzed, as well as for the two crops. The importance factors of exposure scenarios are initially assigned according to available data on the related route. Therefore, scenarios associated with the ingestion route assume values in the range of 7 to 9; those associated with inhalation, from 5 to 9; and those related to dermal adsorption, from 1 to 5 (Rebelo et al. 2020).

It should be noted that water ingestion can occur intentionally, due to lack of information about water potability, or unintentionally, such by as ingestion of microdroplets during sprinkler irrigation (Russo et al. 2020). Both situations are considered in the scenarios of the research.

The exposure routes and scenarios, as well as the assigned importance factors and the respective vulnerabilities of each receptor, are presented in Table 2, for pasture crop.

The exposure routes and scenarios, as well as the assigned importance factors and the respective vulnerabilities of each receptor, are presented in Table 3, for sugarcane.

ISO 16075-1:2015 presents possible barriers to be used in agricultural reuse, as well as the microbiological density reduction and number of equivalent barriers (concept related to a barrier that produces a microbiological reduction to an acceptable level). Based on the ISO 16075-1:2015, US EPA (2012) and WHO (2006) technical standards, the barriers used were defined to determine the damage of each analyzed crop (sugarcane and pasture), as shown in Table 4.

For farmers to use Personal Protective Equipment (PPE), training and capacity building are necessary, in order to make these workers aware of the risks to which they are subjected. Data indicate that farmers do not usually use PPE even for pesticide application, which can lead to health problems such as acute and chronic poisoning (Yarpuz-Bozdogan 2018).

The risk for each receptor and global risk obtained for the application of agricultural reuse in areas around the WWTP in the cities for the analyzed crops are presented in Table 5.

The estimated risk for all receptors is in the acceptable category, with sugarcane and pasture farmers being the most at risk (Table 5). Rebelo et al. (2020), when applying the same methodology of the present research to assess the microbiological risk of the application of agricultural reuse in wineries, also obtained the highest risk for the receptor farmer.

The numerical value of the risk, both by receptor and global, was close to the acceptable risk limit, 7, which could lead to unacceptable risk. In this case, possible minimization measures can be adopted, involving the terms hazard, vulnerability and/or damage.

It is possible to reduce the hazard importance factor by promoting improvements in the reclaimed water treatment system. According to Rebelo et al. (2020), an importance factor for hazard equal to 3 is equivalent to a density of E. coli in the order of 10¹ to 10² CFU/100 mL, which is achievable with secondary treatment associated with disinfection and chlorination (Figure 2).

The adoption of a hazard importance factor of 3 or less results in despicable risk for the same vulnerability configurations (receptors, scenarios and factors) and damage (barriers), as shown in Table 6.

It is possible to obtain despicable risk by changing the damage to values less than 0.43 (for sugarcane and pasture), for the same hazard and vulnerability configurations in Table 5, by adopting more barriers and considering a scale of partial damage (d_i) based on demonstrated evidence of damage to health in the event of a barrier failure.

Table 7 presents the risk for each receptor and global risk altering the damage, and Table 8 presents the details of the changes in the damage, with the adoption of less rigid factors and inclusion of a barrier related to bacteriological decay. Such decay was provided by the adoption of a reservoir to store the water used in agricultural reuse. Despite the modifications in the damage, sugarcane farmers continued with an acceptable risk, since it presents higher vulnerability values.

Table 9 presents a summary of the initial risk estimate by receptor, with the hazard (H), vulnerability (V) and damage (D) values and the modifications in each term. The vulnerability values for each receptor in the term ‘modified vulnerability’ are equivalent to adopting the lowest possible values for the scenario's importance factors. Modifying only the vulnerability does not allow for a despicable risk range (between 0 and 3). Modification of hazard and damage allow for reaching risk estimates for each receptor that are lower than the initial estimate.

Although the microbiological risk was acceptable for all receptors, according to the methodology used, there are other risks in the application of agricultural reuse. Therefore, to reduce the probability of damage to human health, based on the recommendations of the World Health Organization, International Organization for Standardization, and United States Environmental Protection Agency, the adoption of measures is suggested such as: use of a reclaimed water storage reservoir to promote microbiological decay; use of personal protective equipment; conducting training; as far as possible, use of a localized irrigation system; and access restriction during irrigation.

Restricting cattle access to pasture after irrigation, as well as cooking foods made from these cattle, reduces the likelihood of human health risk associated with dairy and meat consumption (WHO 2006).

The maximum values indicated in both non-mandatory WHO and Brazilian guidelines for the coliform standard for water reuse in agriculture are more restrictive than the density present in the application of this study. However, there was an acceptable level of risk for all receptors considered for sugarcane and pasture irrigation, with water reuse, around the WWTP in the cities of Aparecida do Rio Doce, Caçu, Lagoa Santa and Quirinópolis (Goiás, Brazil). For Brazil, the most suitable is the adoption of reclaimed water quality standards according to the fit-for-purpose approach (case by case), supported by risk assessment.

The methodology uses scenarios, receptors, barriers and data on the quality of the reclaimed water, all focused on the specifics of the project. Therefore, it allows for a closer approximation to local reality and facilitates the evaluation of options to achieve the lowest possible risk. In this sense, modifications such as improvements in the wastewater treatment system, used for agricultural reuse, were considered (which reduces the hazard), as well as the adoption of different barrier configurations (which reduces the damage). Consequently, the global risks and those for each receptor, keeping the same scenarios and receptors, went from acceptable to despicable.

Nevertheless, the adoption of preventive measures is suggested, such as restricting the access of animals, population and workers during irrigation, use of personal protective equipment, training, capacity building and modification of the sprinkler irrigation system for a localized system.

Finally, risk assessment enables better system-management, bringing information about the many scenarios for decision-making, in addition to encouraging public confidence in the safe application of water reuse.

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.