Risk Assessment for 2-MIB Odor Complaint in Drinking Water Based on Zero-Inflated Log-Normal Model

Abstract

Based on the most often encountered odor problem caused by 2-methylisoborneol (2-MIB), a quantitative method was developed to determine odor complaint risk by combining odorant concentration probability distribution with complaint threshold distribution, followed by a probabilistic approach. Zero-inflated model was used for handling seasonal 2-MIB concentrations with excess left-censored values for warm and cold seasons by adopting maximum likelihood methods. Complaint threshold values, defined as guidance level of a specific odorant to avoid sensitive population complaints, were fitted with a log-normal distribution. 0.6% complaint threshold values were less than the drinking water quality standard of 10 ng/L of China, illustrating that lowering 2-MIB to below 10 ng/L could effectively protect residents from 2-MIB odor nuisance. As expected, 2-MIB in warm seasons showed a higher mean complaint risk (6.35%) compared with cold seasons (2.88%). 85% of 2-MIB concentrations in warm seasons and 90% in cold seasons would have no complaint risk. Using the developed method, 2-MIB complaint risks in drinking water in five cities were determined. This is the first attempt to assess odor complaint risk in drinking water, which will provide risk managers a new insight on the nonhazard effect of pollutants to further establish water quality standards by constraint of complaint risk.