论文部分内容阅读
渭干河-库车河三角洲绿洲是新疆重要的农业生产区,而大规模的土壤盐渍化始终是制约农业可持续发展的障碍因素。针对干旱区典型绿洲盐渍土盐分动态监测中存在的方法问题,首先确定影响土壤盐渍化形成的各影响因子并提取所需信息,然后利用灰色关联度模型计算各影响因子与土壤盐渍化各影响因子之间的关联度,最后经过多次调整网络结构和参数,建立了基于BP神经网络的表层盐渍土盐分预测模型。结果表明,以地下水位、地下水矿化度、潜在蒸散量、坡度值、土壤电导率、总溶解固体(TDS)、pH值和土地利用类型等8个因素为输入因子,土壤含盐量为输出因子的BP神经网络模型可有效预测盐渍土盐分。本研究可为分析和预测土壤盐渍化动态规律提供一种有效可行的新途径。
Weigan River - Kuqa River Delta is an important agricultural production area in Xinjiang, and large-scale soil salinization is always an obstacle to the sustainable development of agriculture. Aiming at the problems existing in the process of salt monitoring of typical oasis in arid area, we first determine the influential factors that affect the formation of soil salinization and extract the required information. Then we use the gray relational model to calculate the influence factors of soil salinization Finally, after several adjustments of network structure and parameters, a salinity prediction model of surface saline soil based on BP neural network was established. The results show that the soil salt content is the input factor with the groundwater level, groundwater salinity, potential evapotranspiration, slope value, soil conductivity, total dissolved solids (TDS), pH and land use types as input factors The BP neural network model of factors can effectively predict the salinity of saline soils. This study can provide a new and effective way to analyze and forecast the dynamic of soil salinization.