The recently observed high levels of fecal indicators in beach sand confound beach water monitoring efforts. The high levels of fecal indicators may be caused by the loss or the reduced activities of common environmental stresses controlling die-off in the sand. Microcosm experiments were conducted to compare the effects of biotic stresses from autochthonous sand bacteria, protozoa, and viruses on Escherichia coli and Enterococcus faecalis in two tropical beach sands. The inhibition of protozoan activities by cycloheximide did not significantly affect the die-off of E. coli, indicating that protozoan predation played a limited role in beach sand. The contribution from phage infection to E. coli die-off was also negligible. Consequently, autochthonous bacteria were identified as the predominant biotic stress to the die-off of E. coli in beach sand. Subsequent experiments demonstrated that the beach sand had a very low protozoan concentration and low protozoan growth potential when compared with various environmental samples. Co-culturing of E. coli with autochthonous sand bacterial isolates significantly enhanced E. coli die-off. PCR-denaturing gradient gel electrophoresis analysis revealed a complex sand bacterial community, suggesting that bacterial antagonistic effects may be widespread. The study also found that E. faecalis exhibited a much longer survival in beach sand compared with E. coli.

High levels of Escherichia coli were frequently detected in tropical soils in Hawaii, which present important environmental sources of E. coli to water bodies. This study systematically examined E. coli isolates from water and soil of several watersheds in Hawaii and observed high overall genotypic diversity (35.5% unique genotypes). In the Manoa watershed, fewer than 9.3% of the observed E. coli genotypes in water and 6.6% in soil were shared between different sampling sites, suggesting the lack of dominant fecal sources in the watershed. High temporal variability of E. coli genotypes in soil was also observed, which suggests a dynamic E. coli population corresponding with the frequently observed high concentrations in tropical soils. When E. coli genotypes detected from the same sampling events were compared, limited sharing between the soil and water samples was observed in the majority of comparisons (73.5%). However, several comparisons reported up to 33.3% overlap of E. coli genotypes between soil and water, illustrating the potential for soil-water interactions under favorable environmental conditions. In addition, genotype accumulation curves for E. coli from water and soil indicated that the sampling efforts in the Manoa watershed could not exhaust the overall genotypic diversity.Comparisons of E. coli genotypes from other watersheds on Oahu, Hawaii, identified no apparent grouping according to sampling locations. The results of the present study demonstrate the complexity of using E. coli as a fecal indicator bacterium in tropical watersheds and highlight the need to differentiate environmental sources of E. coli from fecal sources in water quality monitoring.

Effects of different land uses on densities of Escherichia coli, enterococci, and Clostridium perfringens in the water and soil of a tropical watershed were investigated. Densities of fecal indicator bacteria (FIBs) in the watershed exhibited a clear land-use dependency in the stream water. Significantly higher concentrations were detected in the urban portion of the stream (417, 420, and 44 CFU 100 mL −1 for E. coli, enterococci, and C. perfringens, respectively) than in the forest portion (54, 32, and 5 CFU 100 mL −1 for E. coli, enterococci, and C. perfringens, respectively). High concentrations of FIBs were also detected in the soil of the watershed with concentration ranges of 603-1,820,000, 69-17,000, and 0-525 CFU 100 g soil −1 for E. coli, enterococci, and C. perfringens, respectively, which however were not affected by the different land uses. Prior cumulative rainfall significantly correlated with concentrations of E. coli and enterococci in the urban stream water (r=0.73-0.87, P<0.05), but not with the alternative FIB C. perfringens. Poor correlations were observed in the forest reach of the stream for all FIBs. Furthermore, the concentration of C. perfringens only correlated strongly and significantly with E. coli and enterococci in stream water (r=0.70-0.82, P<0.05), but not in tropical soil, indicating different survival and transport behaviors.

No abstract

No abstract