论文部分内容阅读
基于分形理论,选取黄土高原南部厢寺川林场地区20个典型样点,每个样点取3个剖面,共60个土样的实测土壤颗粒粒径分布数据,分别应用分形模型、对数正态模型、逻辑生长模型、WEIBULL模型预测土壤颗粒累积百分含量,提出一种预测土壤颗粒粒径分布的分形模型。结果表明,在0.002～0.1mm粒级范围内,分形模型对已知土壤资料的粒级个数和预测粒级的大小等因素并不敏感,具有较高的预测精度和稳定性;与对数正态模型、逻辑生长模型和WEIBULL模型相比,分形模型的总体预测误差最小且未出现大误差数据,可以有效对不同土粒分级标准间土壤质地资料进行转换。
Based on the fractal theory, 20 typical samples were selected from the Xiangsichuanchuan forest area in the southern Loess Plateau. The measured soil particle size distribution data of 60 samples were taken from each sampling point. Fractal models were used and lognormal Model, logistic growth model and WEIBULL model were used to predict the cumulative percentage of soil particles. A fractal model was proposed to predict the particle size distribution of soil particles. The results showed that in the grain size range of 0.002-0.1mm, the fractal model was not sensitive to the number of grain size and the size of the predicted grain size of the known soil data, with high prediction accuracy and stability; Compared with WEIBULL model, the normal prediction model and the logical growth model show that the overall prediction error of fractal model is the smallest and no large error data appear, which can effectively convert soil texture data between different soil classification standards.