道路边际土壤重金属分布格局的神经网络模拟——以现代黄河三角洲为例

在实地采样的基础上,应用MLP神经网络模型和GIS技术,对现代黄河三角洲3类不同道路两侧边际50m带状区域内的土壤重金属含量进行了逐栅格模拟.结果表明:东部和北部的自然保护区以及东北沿海地区道路边际土壤重金属含量较低,西南部人口密集区重金属含量较高.研究区道路边际土壤的Mn、Zn、Cr、Cu、Pb、Cd等重金属平均含量分别达到509.57、61.26、58.55 、32.40、26.07mg/kg和0.27mg/kg,Cd污染较为严重,其余金属含量均低于国家土壤环境质量I级标准.道路对其边际重金属的累积作用明显,但对环境尚未造成广泛和显著的危害.不同道路类型的土壤重金属含量高速公路RG>省道RS>县乡公路RX(除Pb外).在风向、土壤性质和地表植被覆盖等主要因素的影响下,道路边重金属含量在30m以内达到峰值,但因重金属扩散机理的差异,变化规律有所不同.

Simulation of distribution pattern of heavy metals in road verge soils using neural network: a case study of modern Yellow River Delta

Heavy metals concentration was simulated in strip area around three different types of roads on Modern Yellow River Delta (YRD) using GIS aided MLP neural network based on filed samples data. The results showed that the concentrations of heavy metals were relatively lower in the nature reserve areas and coastal region locating in the east and north parts of the modern YRD, but relatively higher in the south and west parts of the area where higher population density were observed. The average metal concentrations of Mn, Zn, Cr, Cu, Pb and Cd in soils beside road verge were 509.57, 61.26, 58.55, 32.40, 26.07mg/kg and 0.27mg/kg respectively. Except the slight pollution of Cd, concentrations of other metals were not exceeding the national environmental quality standard level of soils. This study revealed that the road in the study area played an important role in heavy metal accumulation, but did not bring broad and notable environment contamination yet. For different types of roads, the concentrations of the tested metals were in the order of Express Way (RG)> Provincial Road (RS)> Rural Road (RX), except Pb. The concentrations of heavy metals besides road verge were reached at peak value within 30m, and the distribution was mainly dependent on many factors such as wind direction, soil properties and vegetation cover. The different changing patterns of heavy metals were observed in individual metal elements because of their different diffusion mechanism.