BackgroundMore than 60,000,000 Bangladeshis are drinking water with unsafe concentrations of one or more elements.ObjectivesOur aims in this study were to evaluate and improve the drinking water testing and treatment plans for western Bangladesh.MethodsWe sampled groundwater from four neighborhoods in western Bangladesh to determine the distributions of arsenic, boron, barium, chromium, iron, manganese, molybdenum, nickel, lead, antimony, selenium, uranium, and zinc, and to determine pH.ResultsThe percentages of tube wells that had concentrations exceeding World Health Organization (WHO) health-based drinking water guidelines were 78% for Mn, 48% for U, 33% for As, 1% for Pb, 1% for Ni, and 1% for Cr. Individual tube wells often had unsafe concentrations of both Mn and As or both Mn and U. They seldom had unsafe concentrations of both As and U.ConclusionsThese results suggest that the ongoing program of identifying safe drinking water supplies by testing every tube well for As only will not ensure safe concentrations of Mn, U, Pb, Ni, Cr, and possibly other elements. To maximize efficiency, drinking water testing in Bangladesh should be completed in three steps: 1) all tube wells must be sampled and tested for As; 2) if a sample meets the WHO guideline for As, then it should be retested for Mn and U; 3) if a sample meets the WHO guidelines for As, Mn, and U, then it should be retested for B, Ba, Cr, Mo, Ni, and Pb. All safe tube wells should be considered for use as public drinking water supplies.

A procedure has been developed for calculating permeability (k) from the Kozeny-Carman equation, a procedure that links ideas from percolation theory with the ideas of Koltermann and Gorelick (1995) and Esselburn et al. (2011). The approach focuses on the proportion of coarser pores that are occupied by finer sediments relative to a percolation threshold proportion (ω(c)). If the proportion occupied is below ω(c), then the unoccupied coarser pores percolate. Otherwise they do not percolate. Following the ideas of Koltermann and Gorelick (1995), the effective grain-size term in the Kozeny-Carman equation is calculated using the geometric mean if the unoccupied coarse pores percolate, and using the harmonic mean if otherwise. Following ideas of Esselburn et al. (2011), this approach is implemented by evaluating the potential for grains in each size category to occupy pores among sediment of each larger-size category. Application of these ideas to physical sediment models for sands and gravels, which have known k, indicates that a threshold does indeed exist. Results also suggest that the Kozeny-Carman equation is robust and gives representative values for k, even though ω(c) is not precisely known.