Using fuzzy clustering algorithms to describe the distribution of trace elements in arable calcareous soils in northwest Iran
2013
Nourzadeh, M. | Hashemy, S. M. | Martín, Juan Antonio | Bahrami, H. A. | Moshashaei, S. | Martín, Juan Antonio [0000-0002-3207-1319]
Accumulation of trace elements in arable soils is an important global hazard worldwide. In this research, the available content of Zn, Fe, B, Co, Cu, Mn, Mo and other soil parameters (pH, organic carbon content, carbonates and electrical conductivity) were analysed in northwest Iran. Concentration levels of trace elements were relatively low in areas with high pH values and low organic matter content, and only the Mo value exceeded the reference threshold. Based on the correlation among the elements, two datasets were produced. The first consists of Fe and Mn data, while the second contains Zn, B, Co, Cu and Mo data. Two fuzzy clustering approaches, Fuzzy C-means (FCM) and Gustafson-Kessel (GK), were applied for clustering both datasets. Multiple accumulation of trace elements was investigated from the clustering results and then visualized in spatial regionalization maps. The fuzzy clustering evaluating indices showed that the GK method was more appropriate than FCM for clustering datasets. The results revealed that the first and second datasets were divided into seven and six clusters, respectively. Fuzzy clustering analyses combined with geostatistical methods were used to map the spatial variability of each cluster. This method enabled the monitoring of multiple metal accumulation in large agricultural soils. © 2013 Copyright Taylor and Francis Group, LLC.
显示更多 [+] 显示较少 [-]