Source Apportionment and Geographic Distribution of Heavy Metals and as in Soils and Vegetables Using Kriging Interpolation and Positive Matrix Factorization Analysis

Abstract

Food security and cultivated land utilization can be seriously affected by heavy metal (HM) pollution of the soil. Therefore, identifying the pollution sources of farmland is the way to control soil pollution and enhance soil quality effectively. In this research, 95 surface soil samples, 34 vegetable samples, 27 irrigation water samples, and 20 fertilizer samples were collected from the Wuqing District of Tianjin City, China and was used to determine their HMs accumulation and potential ecological risks. Then, kriging interpolation and positive matrix factorization (PMF) were utilized to identify the sources of soil HMs. The results indicated that soil HMs in the study area were contaminated at a medium level, but that the pollution of Cd was more severe, and the Cd content in vegetables was slightly higher than the permissible threshold (0.02 mg·kg−1). Furthermore, a non-homogeneous distribution was observed, with higher concentrations of HM contaminants concentrated in the southwest of the study area, where many metal manufacturing industries are located. Our results suggest that the Cd originated from industrial activity; As and Pb from agricultural practices; Ni, Cu, Cr, and As mainly from natural sources; Zn and Cu from organic fertilizer; Pb and Cd mainly from traffic discharge; and Cr, Ni, and Pb from sewage irrigation. Obviously, the accumulation of soil HMs in the study area could be mainly attributed to industrial activities, implying the need for implementation of government strategies to reduce industrial point-source pollution.

Spatial distribution and source identification of metal contaminants in the surface soil of Matehuala, Mexico based on positive matrix factorization model and GIS techniques

“…The different techniques were utilised to determine the distinctive distribution patterns and sources of metal contamination in soils, including geographic information system (GIS), positive matrix factorization (PMF), and principal component analysis (PCA) (27)(28)(29)(30)(31). To identify the source contributions, standard source apportionment procedures (e.g., PCA and PMF) and contamination monitoring techniques (e.g., contamination indices and ecological risk factors) are often implemented (32-36) and estimate the risk of metal contamination in soils, accordingly (37)(38)(39)(40).…”

“…The positive matrix factorization (PMF) model is a mathematical matrix decomposition approach that can manage insufficient and imprecise data by applying nonnegative constraint criteria. This model is a common receptor model which has been frequently utilised for the identification of the source of metal contamination in the atmosphere, sediment, and soil (31,57,58). However, the PMF model is empirical in identifying sources of contamination and assumes contaminant spreads as a linear process (56,59).…”

Estimating Source Apportionment of Heavy Metals Contamination in Surface Soil Based on a Positive Matrix Factorization (PMF) Model around Cerrito Blanco in San Luis Potosi, Mexico

“…Heavy metals accumulate in the soil through atmospheric deposition, sedimentation, sewage irrigation, as well as other routes as a result of excessive industrial and agricultural activities [1,2]. However, there are several geological and anthropogenic sources of heavy metal contaminants, including air deposition, industrial discharges, fertiliser application, pesticide usage, etc., that might contribute to their presence in the soil in urban-rural peripheral zones [3,4]. Therefore, it is essential to evaluate the heavy metal contamination along with necessary toxicological assessment.…”

“…Additionally, the anthropogenic sources of soil heavy metals include mining, agriculture, vehicle emissions, ore deposits, industrial emissions, coal burning, atmospheric deposition etc. [4,8,10]. To determine the source identifications and the spatial variations of heavy metals in soils, multivariate statistics are implemented in combination with geostatistical techniques [11,12].…”

“…There are also many receptor models which are used to identify the sources of heavy metals. The most prominent and widely used receptor models are principal component analysis/absolute principal component scores (PCA/APCS), positive matrix factorization (PMF), and UNMIX [4,13,14]. In the early 1990s, source identification of pollutants in the atmosphere was done using positive matrix factorization (PMF) [15].…”