论文部分内容阅读
在对石家庄市边缘带进行划分的基础上,利用BP神经网络模型并结合地统计学方法研究了石家庄市边缘带土壤养分的空间分布特征。结果表明:BP神经网络模型能够智能地学习各个采样点的空间位置与该点各种土壤养分含量之间的映射关系,并能稳健地对不同地理位置上的土壤养分含量进行预测。预测后的有机质、全氮、速效磷和速效钾变异函数的理论模型均符合球状模型。从养分分布图看,石家庄市边缘带有机质、全氮、速效磷和速效钾含量具有明显的斑块状分布特点,各种土壤养分高值区多分布在石家庄市西部和南部。4种养分含量均呈现从西向东、从南向北递减的趋势。通过引入土地景观格局指数和土壤养分综合得分两个指标,建立回归模型,并对土壤养分含量与土地利用格局进行了研究。结果显示:土地景观格局指数与土壤养分的主成分综合得分拟合模型为y=1.014xexp(x/1.794)+2.152,两者之间存在明显的指数增长趋势。景观格局指数越高,土地利用类型越复杂,土壤养分含量相应偏高,反之亦然。
Based on the division of the marginal zone of Shijiazhuang City, the spatial distribution characteristics of soil nutrients in the marginal zone of Shijiazhuang City were studied by BP neural network model and geo-statistics method. The results show that the BP neural network model can intelligently learn the mapping relationship between the spatial positions of various sampling points and various soil nutrient contents at this point, and can predict the soil nutrient contents in different geographical locations steadily. The theoretical models of the predicted variances of organic matter, total nitrogen, available phosphorus and available potassium all fit the spherical model. From the nutrient distribution map, the edge of Shijiazhuang City with organic matter, total nitrogen, available phosphorus and available potassium content has obvious patch-like distribution characteristics, a variety of high-value soil nutrients are distributed in the western and southern Shijiazhuang City. The four nutrient contents showed a decreasing trend from west to east and from south to north. Through the introduction of two indexes of land landscape pattern index and soil nutrient comprehensive score, the regression model was established, and the soil nutrient content and land use pattern were studied. The results showed that the fitting model of the composite score of land landscape pattern index and soil nutrients was y = 1.014xexp (x / 1.794) +2.152, with a clear trend of exponential growth. The higher the landscape pattern index, the more complex the land use types, the higher the soil nutrient content, and vice versa.