Prediction of pesticide residues in agricultural products based on time series model in Chengdu, China

Abstract

Pesticides are chemicals that can improve the efficiency of agricultural production, but also cause harm to human health and the environment. Besides effective supervision means, prevention work is also indispensable for pesticide residue safety in agricultural products. The Shuangliu and Pidu districts of Chengdu, capital city of Sichuan Province, China, were selected as the focus areas. Pesticide residue levels in leafy and starchy vegetables were measured for 15 months, analyzed with paired-sample t-Tests to construct four ARIMA time series models. The results showed that pesticide residue levels of different types of agricultural products in the same area as well as those of the same agricultural products in different areas differed within the same time interval. Meanwhile, the pesticide residue levels of different agricultural products and areas showed distinct seasonal characteristics and variations. The ARIMA models were effective for short-Term forecasting of agricultural pesticide residue levels. They could be used in related fields to predict crop pesticide residue levels pre-emptively based on actual usage patterns, crop type, season, and other parameters. The findings in this study may help government and traceability agencies releasing early warnings of potential agricultural product pesticide contamination. It may also help mitigate the risk of crop and environmental contamination as well as consumer health endangerment.