USING SURVIVAL ANALYSIS TO ESTIMATE STATISTICS OF CENSORED ENVIRONMENTAL OBSERVATIONS

It has become increasingly common for water, soil, air and waste streams' environmental monitoring data to include censored values that reflect more than one detection limit for a single analyte. To assess censored (environmental) data continues to be a topic of regulatory authorities, environ-statisticians and characterization and remediation engineers. Chemicals, mainly trace elements contaminated in groundwater, soils or waste streams, need to be quantified for regulatory compliance and environmental impact assessment, and thus play a role in decision-making. However, single or multiple censoring related to analytes result in incomplete data and become an obstacle to obtaining general statistics of data distributions. Survival analysis is used to evaluate analyte observations composed of both quantified and censored data. The originality of survival analysis addressing right censored data is referred to in the case of left censored environmental data. Summary statistics, i.e. mean, median, percentile, cumulative distribution function and survival function, are estimated, which offer a baseline for characterizing contaminant carriers.