Impact of population and latrines on fecal contamination of ponds in rural Bangladesh

Abstract

A majority of households in Bangladesh rely on pond water for hygiene. Exposure to pond water fecal contamination could therefore still contribute to diarrheal disease despite the installation of numerous tubewells for drinking. The objectives of this study are to determine the predominant sources (human or livestock) of fecal pollution in ponds and examine the association between local population, latrine density, latrine quality and concentrations of fecal bacteria and pathogens in pond water. Forty-three ponds were analyzed for E. coli using culture-based methods and E. coli, Bacteroidales and adenovirus using quantitative PCR. Population and sanitation spatial data were collected and measured against pond fecal contamination. Humans were the dominant source of fecal contamination in 79% of the ponds according to Bacteroidales measurements. Ponds directly receiving latrine effluent had the highest concentrations of fecal indicator bacteria (up to 10⁶ Most Probable Number (MPN) of culturable E. coli per 100 mL). Concentrations of fecal indicator bacteria correlated with population surveyed within a distance of 30-70 m (p<0.05) and total latrines surveyed within 50-70 m (p<0.05). Unsanitary latrines (visible effluent or open pits) within the pond drainage basin were also significantly correlated to fecal indicator concentrations (p<0.05). Water in the vast majority of the surveyed ponds contained unsafe levels of fecal contamination attributable primarily to unsanitary latrines, and to lesser extent, to sanitary latrines and cattle. Since the majority of fecal pollution is derived from human waste, continued use of pond water could help explain the persistence of diarrheal disease in rural South Asia.

Figure 1

Concentration of cultured E. coli in each pond classified by pond type with locations of sanitary and unsanitary latrines. IKONOS satellite image taken of the entire region of Araihazar at 1 m resolution (van Geen et al. 2003).

Figure 2

Dissolved oxygen, electrical conductivity, temperature and concentrations (MPN or copies/100 mL) of three fecal markers in water, from three types of ponds at Site K. The center line represents the median, upper and lower bounds of the box are the 75^th and 25^th percentile and whiskers represent the extent of the data. Outliers are represented by dots. The number of ponds were 43, consisting of 11 fish/bathing, 16 latrine and 16 ponds with no defined use in each category. The geometric mean detection limit for molecular assays was 8,374 copies/100 mL and is indicated by the dotted line.

Figure 3

Observed and predicted E. coli and total Bacteroidales as a function of mE. coli in 43 pond water samples. One non-detect occurred with culturable E. coli and 3 non-detects occurred for mE. coli, but not in the same samples. Total Bacteroidales was detected in every pond water sample. Error bars represent 95 % analytical confidence intervals. Predictive equation is the result of fitting a linear regression model (y = mx + b) to log₁₀-transformed concentrations with R². The 1:1 line is shown for comparison (y=x).

Figure 4

Relative amounts of fecal contamination from Human or Bovine sources. Blue circles represent ponds where Human or Bovine Bacteroidales were not detected.