基于模拟退火算法的土壤样点布局和土壤景观模型优化

发布时间：2018-09-09 08:24

【摘要】：本研究利用多重线性回归方程,以地形因子为预测变量,构建关于土壤有机质的土壤景观模型,并以西南山地丘陵区的一块面积为2 km2的汇水盆地为研究区,对该区域的土壤有机质空间分布进行预测。在此基础之上,探讨最少可用多少个点来预测土壤有机质的空间分布,并使之预测精度不低于原始集合的精度;同时,找出最优土壤样点布局,确定不同地形部位的取样单元,使之预测精度最高。研究结果表明:在预测误差最小化的情况下,最少可用7个优化的样点就可以代替原始200个采样点,且优化的样点数为124时,模型预测土壤有机质空间分布的精度最高。优化后的土壤景观模型的拟合度比原始模型提高了3.28%,MAE降低了5.3%,RMSE降低了3.94%。
[Abstract]:In this study, a soil landscape model of soil organic matter was constructed by using multiple linear regression equations and topographic factors as predictive variables, and a catchment basin with an area of 2 km2 in the mountainous and hilly region of southwest China was used as the study area. The spatial distribution of soil organic matter in this area was predicted. On the basis of this, we discuss how many points can be used to predict the spatial distribution of soil organic matter, and make the prediction accuracy not lower than that of the original collection. At the same time, we can find out the optimal soil sample distribution and determine the sampling units in different topographic areas. The prediction accuracy is the highest. The results show that at least 7 optimized samples can be used to replace the original 200 sampling points when the prediction error is minimized and the model has the highest precision in predicting the spatial distribution of soil organic matter when the optimum sample number is 124. Compared with the original model, the fitting degree of the optimized soil landscape model was increased by 3.28% and decreased by 5.3% and 3.94% by RMSE.
【作者单位】：西南大学资源环境学院;中国热带农业科学院橡胶研究所;西南大学计算机与信息科学学院;
【基金】：国家科技支撑计划课题(2008BADA4B10) 中央高校基本科研业务费专项(XDJK2016D041)资助
【分类号】：S151.9

【相似文献】