论文部分内容阅读
土壤有机质是农田土壤质量的重要指标,精确地预测其空间分布是必要的.然而,大多数的研究仅仅是在一个较小的范围内,较大的采样密度下的预测.低山丘陵区农田土壤有机质受多种环境因素的影响,为实现较小采样密度下大区域土壤有机质精确预测,本研究应用预测土壤制图技术,在西南地区的一个面积为3 200km2的典型低山丘陵区,收集和分析了2 000个土壤样点(1 600个建模点,400个检验点),辅助以环境因子(地形因子、植被指数和土地利用),通过逐步回归构建了区域土壤有机质与环境因子的预测模型(R2=0.453,F=440.3,p<0.05),对土壤有机质的含量进行了预测性制图.利用验证集对线性回归模型预测结果与克里格插值结果做了比较,显示低山丘陵区较小采样密度下,由于线性回归模型考虑了环境因素的空间变异特征,其预测结果优于常用的普通克里格插值模型.
Soil organic matter is an important indicator of farmland soil quality and it is necessary to accurately predict its spatial distribution.However, most of the researches are only based on a smaller range and larger sampling density.Forest Soil organic matter is affected by many environmental factors. In order to achieve accurate prediction of soil organic matter in large sampling areas with small sampling densities, this study applied predictive soil mapping technology to a typical hilly area of 3 200 km2 in the southwest of China. A total of 2 000 soil samples (1 600 modeling points and 400 checkpoints) were analyzed to assist the prediction of soil organic matter and environmental factors in the region by stepwise regression using environmental factors (topographical factors, vegetation indices and land use) (R2 = 0.453, F = 440.3, p <0.05), the content of organic matter in the soil was predicted by using the validation set.The linear regression model was compared with the results of Kriging interpolation.It shows that the hilly region At smaller sampling densities, the linear regression model takes into account the spatial variability of environmental factors, and its prediction is better than the common kriging interpolation model.