Groundwater bacteriological quality assessment: impact of urbanization and agricultural activity

Groundwater from boreholes is the major source of bottled water in Algeria. The aim of this study is to determine the bacteriological quality of groundwater that serves bottled water production. A total of 73 groundwater boreholes were sampled and analyzed for the required bacteriological parameters. The analysis was performed in accordance to ISO standards methods. There should be no bacteria growth for each bacteriological parameter to qualify the groundwater of good bacteriological quality. The bacteriological analysis highlighted that 37 of the 73 groundwater samples (51%) were of poor bacteriological quality while 36 of them (49%) were of good bacteriological quality. Total coliforms and E. coli were the major sources of contamination with respectively 35 and 24 contaminated samples, followed in order by Pseudomonas aeruginosa, enterococci and sulfite reducing anaerobic bacteria spores with respectively 8, 7 and 2 contaminated samples. Bacteriological quality was strongly and negatively correlated with urbanization and/or agricultural activity parameter (r1⁄4 0.454). The performed logistic regression model showed that the presence of urbanization or agricultural activity multiplies significantly (P< 0.001) the risk by 7 of being a poor bacteriological quality groundwater. These findings are useful to avoid drill costs and to take the best strategy to protect groundwaters.


INTRODUCTION
Algerian bottled water demand is increasing (Hazzab ).
The total production comes from groundwater. Its total potential is estimated 2.7 billion m³ in the Atlas northern region and 5 billion m³ in the southern region of the Sahara (Chabour et al. ). Groundwater provides 63% of the total water demand in the Northern region and 96% of the water demand in the Sahara region (FAO ).
Taking into consideration its impact on public health and the economic sector, it is important to assess and monitor the groundwater quality, with focus on the bacteriological quality for the prevention against waterborne disease.
Groundwater destined for bottling is subject to rigorous bacteriological control before putting it in the market (Ministry of Water Resources ). The cost of drilling being important (World Bank ), the determination of the factors influencing the bacteriological quality of groundwater is useful in the perspective of avoiding financial waste and providing more rationale to the feasibility studies.
National regulations define the groundwater bacteriological quality parameters and the sample size that should be analyzed for each parameter (Ministry of Water Resources ; Ministry of Trade ). In this regard, we are generally seeking for fecal contamination indicators. The water is considered to be of good bacteriological quality when there is no growth of bacteria for each evaluated parameter.
Bacteriological pollution of groundwater is generally due to the presence of fecal contamination bacteria indicators, originating mostly from humans or animals (Koffi-Nevry et al. ), through intense urbanization, that produces domestic effluents and agricultural activity via animal husbandry or the use of manure and slurry as natural fertilizer (Seiler et al. ; Esterhuizen et al. ). Several previous studies were carried out to assess the bacteriological and physicochemical quality of groundwater in few regions of Algeria, revealing considerable pollution and concluding on assumptions about the origin of such pollution (Fehdi et al. ; Gueroui et al. ; Benouara et al. ). Sampling size and localization of these studies suffered from several limitations pertaining mainly to small sample sizes and focusing only on one region. Thus, in the current study, an intense effort has been made to collect more samples over many provinces of Algeria, in order to strengthen the results. A study carried out in the city of Bechar did not demonstrate a significant impact of urban domestic effluents on the groundwater bacteriological quality (Kabour et al. ). The previous studies highlighted the lack of agricultural activity and urbanization impact on the groundwater bacteriological quality in Algeria and paved the path to this study towards the examination of this question. We collected multiple samples from several boreholes in different provinces of Algeria, carried out a bacteriological analysis. We also constructed a dataset on factors that may influence the bacteriological quality of groundwater, mainly the depth of the borehole, the temperature of water at source, and anthropic activities such as the presence or absence of nearby agricultural or urban activity from the borehole site.
This study aims to assess the bacteriological quality of groundwater in Algeria and to determine the impact of factors that may influence it, in terms of agricultural activity and urbanization.  supplementary information, Table S1). The boreholes were purged by a pump through stainless steel pipe. A total of 73 groundwater samples were collected from the study area during early morning hours (07:00-09:00 h). The person collecting the samples should wear sterile gloves.

Study area and study design
The water should be left to run from the borehole for a few hours by turning on the tap before sampling. The steel tap should be flamed with alcohol and gas torch then samples were collected in sterile polyethylene terephthalate bottles of 500 mL and transported in cooler. They were stored under (5 ± 3) C and were sent to the laboratory for examination within maximum 12 h from their collection.

Laboratory analysis of samples
Groundwater bacteriological analysis was performed using the membrane filtration technique and the standing tube technique to assess the bacteriological status of the groundwater for the required bacteriological parameters, according to the standard methods presented in Table 1.
The membrane filtration technique consisted of passing 250 mL of groundwater samples in sterile funnels using vacuum filtration ramp for each parameter through a cellulosic membrane filter with pores of uniform diameter equal to 0.45 μm. After filtration, this membrane filter was inoculated in a Petri dish containing a standard culture medium specific for each bacteriological parameter. We used TTC tergitol-7 agar for total coliforms, Slanetz and Bartley medium for enterococci and cetrimide agar for Pseudomonas aeruginosa. At the end of the operation, the Petri dishes were placed in the incubators set at 37 C.
The results were collected by counting the colonies 48 h after incubation, except total coliforms that we could collect from 24 to 48 h after incubation. Total coliforms were confirmed by oxidase test which should be negative. E. coli was confirmed from total coliforms with indole production at an elevated incubation temperature (44 C) and with a positive β-glucuronidase test. Enterococci were confirmed by esculin hydrolysis in the presence of bile. The membrane containing suspicious colonies was inoculated in a Petri dish of bile-esculin-azide agar and incubated 2 h at 44 C, esculinase positive colonies gave a brown-black halo. Colonies producing blue-green pigment (pyocyanin) in cetrimide were directly counted as Pseudomonas aeruginosa. Those

RESULTS AND DISCUSSION
In situ data The analysis of this data showed an average score for bore-

Bacteriological quality assessment
The results of the bacteriological analysis are presented in Table 3. They showed that overall the half of the sampled groundwater boreholes did not meet the good bacteriologi-   by total coliforms (Figure 6). This means that total coliforms were the most important source of contamination in our samples and the first one which makes them of poor quality.
Total coliforms were also the most common microbiological contaminant in the Pennsylvania study (Swistock et al. ). their presence in the environment can have other reasons.
These spores can survive in groundwater much longer than coliforms, due to their longevity; they are mainly able to indicate intermittent or remote contamination, and the vulnerability of the aquifer. Therefore, the pollution origin is very far from the borehole in time and space (WHO ).

Correlation matrix
Correlation matrix was performed on all the studied parameters (Table 4). We focused on strong correlations that may bring some information from our dataset. Bacteriological quality was highly and negatively correlated with total coliforms (r ¼ À0.947) and E. coli (r ¼ À0.690). Thus, total coliforms and E. coli were the most influencing groundwater bacteriological quality in our work as it was shown above. Bacteriological quality was also strongly and negatively correlated with urbanization and/or agricultural activity parameter (r ¼ À0.454). This supposes that this last parameter may negatively affect groundwater bacteriological quality in the studied area. Therefore, we explored this association by a logistic regression (regression results are presented in a subsequent section). Sulfite reducing anaerobic bacteria spores was highly and positively correlated with enterococci (r ¼ 0.465). This is possibly inherent to the fact stipulating that they may come collectively from fecal contamination and being resistant than the other groundwater parameters or their simultaneous absence or presence. It is important to note that the presence of them in our samples was rare, so we need further data collection to demonstrate such fact. Finally, groundwater temperature at source was significantly and positively correlated with borehole depth (r ¼ 0.352) because the temperature of earth is rising towards its center; the greatest depth induces high groundwater temperatures, except boreholes less than 10 m that are affected by ambient temperature of the atmosphere (Pedersen et al. ). In our study, all our boreholes were more than 10 m deeper. Therefore, the correlation was relatively preserved.
(2): The binary value of urbanization and/or agricultural activity variable that could take 0 or 1.

CONCLUSION
The results of this study indicate that almost half of groundwater samples did not satisfy the regulatory bacteriological quality criteria, specifically for total coliforms followed in