Analysis of Big Data Obtained from Continuous Measurement of 22 Malodor-producing Substances in Korean Cattle Farms

Abstract

Objectives : This research collected continuous (20-minute interval) measurement data of 22 compounds from cattle farms during a year in order to understand malodor emission characteristics and trends and also big data analysis was performed to analyze the relationships between malodor concentration and environmental factors such as wind direction and speed, temperature, and humidity.Methods : The measured substances were 22 compounds defined as ‘designated malodor-producing substances’ by Korean Ministry of Environment. Detection and concentration measurement were performed by gas chromatography with two kinds of columns (VB-WAX, and VB-1), and detectors (FID and PID). Sampling was carried out with pumping every 2 or 5 minute and thermal desorption was followed by chromatograph injection. Data obtained from analysis were stored with environmental variables such as temperature, humidity, wind speed, and wind direction from 1 August 2022 to 31 July 2023. Total number of measurement data was 26,280.Results and Discussion : Correlation analysis showed that temperature was correlated with dimethyl disulfide, and iso-valeraldehyde, and humidity was correlated with toluene, iso-butylalcohol, and trimethylamine. Six principal components explained 60.7% of total variance and two principal components, PC 1 and PC 2, were selected and analyzed. PC 1 was strongly related with volatile organic compounds and PC 2 was related with acid compounds. Together with bipolar plot results, PC 1 was defined as volatile properties and PC 2 as acidic characteristics, respectively. Multiple regression with principal components as dependent variables and environmental variables as independent showed relatively small value of adjusted R2 except PC 1 of 0.539.Conclusion : Malodorous compounds tended to diffuse and move together with similar molecular characteristics which made the detected concentration similar in patterns. Though PC 1 had correlation with environmental variables through multiple regression analysis, it can be explained as chemical compound’s characteristics depends on temperature, humidity, and wind.