Abstract:
Countries suffering from water shortage problems use the intermittent mode of supply
where water is pumped for few hours once or twice a week, which forces householder to
store water in tanks located on rooftops or in reservoirs beneath the house. Water stored
for several days may rapidly deteriorate imposing health risks on consumers. The first part of this study aimed at the examination of the microbial populations present
in different household water storage tanks, and to examine the extent and the
phenomenon of this microbial regrowth and determine the factors involved in it.
The heterotrophic plate count (HPq of the household water storage tanks in Byblos city,
Lebanon ranged from 5 CFU/ mL to 4xl04 CFU/ mL. A significant difference was not
detected in the total bacterial count in the samples collected from the different types of
tanks (black polyethylene, fiberglass, asbestos (etemite) and galvanized cast iron).
Pseudormrns (P. aeru[!jna;a, 26.2%) was the dominant type observed, while Bacillus
(B. amJdiqu{aciens, 9.63%) was the next dominant organism isolated from the storage
tanks. Opportunistic pathogens such as P. aempjnaa (2 of 7 samples), E. sakazakii (2 of 7),
E. a[§!}onrrans (1 of 7) and A eronvnas sdJri£t (1 of 7), in addition to the dominant bacterium
Burkhdderia cepacia (5 of 7) were detected in the drinking water samples collected from
Byblos city. The presence of such opportunistic pathogens could be attributed to
inefficient chlorination. Thus the city drinking water is not of acceptable microbiological
condition after reaching consumers due to the considerable occurrence of such
microorganisms. However, P. aerugjnaa and coliforms were not detected in the UV treated
drinking water samples. The effect of water retention time on bacterial regrowth pattern in the chlorinated city
fiberglass tank was studied. Results obtained indicated bacterial regrowth with the
dominant species being Pseudormnas spp. Bacterial exponential growth occurred during the
first 5 days with an HPC increase of 1400 fold ( 20 CFU/rnL to 28000 CFU/rnL) ,
compared to an increase of only 1-fold during the period between 5 and 7 days storage
time. In the UV treated water 9 to 16-fold increase in the HPC level was detected after 2
and 4 days of storage in the PE and cast iron tanks, respectively. In the UV treated water storage tanks, data showed that the frequencies of occurrence of
gram-negative bacteria were 54% in the cast iron, 83% in the PE and 75% in the influent
water. In the chlorinated city water tanks, the frequencies of occurrence of gram-negative
bacteria were 86% in the cast iron, 50% in the PE, and 87% in the influent water. Chlorination experiments performed on the water stored in the PE tanks showed that
maintaining a free residual chlorine 2: 0.25 mg/L was enough to control the HPc, while a
271-fold decrease (27100 CFU/rnL to 100 CFU/rnL) of HPC bacteria occurred at a free
chlorine of 0.11 mg/L. The gram-positive OJrynebaaeriumutriabilis survived free [02] :::; 0.17
mg/L in the water of PE tanks. On the other hand, maintaining a free residual chlorine
2: 0.35 mg/L was enough to control HPC levels in the cast iron tanks. In contrast to the
PE tanks, some bacteria survived in the presence of 0.26 mg/L and 0.19 mg/L of free
chlorine in the cast iron tanks. At 0.19 mg/L, endospore-forming gram-positive Bacilli were the dominant types detected, while at 0.26 mg/L an organism suspected to be
Myr1tuterium was detected.
This study showed that 91.2% of the gram-negative bacterial isolates could be identified by
the Biolog system, whereas only 82.5% were identified by the API system. The percentage
of isolates that were assigned the same identity (genus & species) by both systems was
24.6%. Moreover 33.3% were not similar at the genus level, while 15.8% were not similar
at the species leveL Difficulties encountered in the identification of these bacteria were
possibly due to the oligotrophic nature of these isolates. However, 89% of the grampositive
bacterial isolates were identified using the Biolog system. Identification of 31 isolates belonging to Pseudarmnas however, revealed better agreement
in the results at both the genus and species leveL The percentage of isolates that were
assigned the same identity by both systems was 45.2%. Differences were seen between
both systems, where 9.68% and 35.5% of the isolates using the API were assigned to
different species or genera, respectively. However, 6.45% were identified as Pseudarmnas by
the Biolog and not identified by the API, and 3.22% were identified as Pseudarmnas by the
API and not identified by the Biolog. Furthermore, among the coliform group isolated and
belonging to 2 genera CitrrlJaaer and E~, both systems showed 100% similarity at
the genus level, and no similarities at the species level.