This study evaluated the effect of population estimation on the calculation of drug biomarker consumption using wastewater-based epidemiology. Population estimates using mobile phone data, census data, and wastewater quality parameters, such as biological oxygen demand (BOD), total nitrogen (TN), and total phosphorus (TP), were evaluated in six different wastewater treatment plant catchment areas of Busan Metropolitan City, South Korea. The population based on mobile phone data was affected by the patterns of non-resident population movements in each area. The population-normalized daily loads (PNDLs) of methamphetamine were compared according to the different population results. The PNDLs using the population based on mobile phone data (PNDLMobile) was 5.87-27.0 mg/d/1000 people. The PNDLMobile values were notably different from the PNDLs using wastewater quality parameters (PNDLWastewater) (PNDLWastewater/PNDLMobile: 51-148 %, mean 93 %, relative standard deviation (RSD) 36 %), indicating the unsuitability of population estimation using BOD, TN, and TP. In areas with a large concentration of workplaces, the PNDLs using census data (PNDLCensus) differed from the PNDLMobile values (PNDLCensus/PNDLMobile: 57-124 %, mean 94 %, RSD 27 %), whereas other areas showed similar values for PNDLCensus and PNDLMobile (PNDLCensus/PNDLMobile: 95-108 %, mean 102 %, RSD 4.2 %). In particular, the total population estimates of the six survey areas using census data were approximately the same as those based on mobile phone data (RSD: 0.8 %), indicating a decrease in the influence of the non-residential active population in the entire metropolitan city.
Keywords: Cellphone data; Census data; Hydrochemical parameters; Population estimation; Wastewater-based epidemiology.
Copyright © 2022 Elsevier B.V. All rights reserved.