| 摘要: |
| 目的 数字高程模型（Digital Elevation Model， DEM）的分辨率大小决定区域地貌形态特征的表达程度，但目前DEM对于局部地形变异大的平原地区土壤有机碳制图精度的影响尚不明确。方法 文章以江苏北部耕地区域为例，利用随机森林算法，建立不同分辨率DEM下的地形、气候、植被、土壤、成土母质与土壤有机碳之间的关系模型，分析DEM分辨率对局部地形变异大的平原区土壤有机碳制图精度的影响程度。结果 （1）不同DEM分辨率下，地形、气候、植被、土壤与成土母质5类因子对土壤有机碳含量的影响程度依次降低；（2）当DEM分辨率在大于150 m的范围内变化时，环境特征因DEM的“概化”而逐步缺失细节信息，从而导致土壤有机碳含量预测精度随分辨率的增加而降低；（3）地形是苏北耕地区域土壤有机碳空间分布差异的最重要影响因子。结论 DEM分辨率在60－90 m范围内时，整体平坦但局部地形变异程度较大的苏北地区土壤有机碳制图效果最佳。该研究为进一步提高我国相似地貌类型区土壤有机碳制图精度提供了理论依据。 |
| 关键词: 土壤有机碳含量 随机森林 地形 DEM分辨率 数字土壤制图 |
| DOI：10.7621/cjarrp.1005-9121.20220726 |
| 分类号:S159.2 |
| 基金项目:福建省自然基金“多制图尺度下福建省耕地土壤有机碳不确定性研究”（2020J05027）；国家自然科学基金“亚热带水田和旱地土壤有机碳模拟的尺度效应差异机理研究”（41971050） |
|
| RESEARCH ON EFFECT OF DEM RESOLUTION ON THE DIGITAL SOIL ORGANIC CARBON MAPPING FOR CULTIVATED AREA IN THE NORTH OF JIANGSU PROVINCE |
|
Wu Ting1,2, Wu Qihang1,2, Cao Wenqi1,2, Chen Jingyan1,2, Liu Shaogui3, Yu Dongsheng4, Shi Xuezheng4, Xing Shihe1,2, Zhang Liming2
|
|
1.College of Resources and Environment, Fujian Agriculture and Forestry University, Fuzhou 350002, Fujian, China;2.Fujian Provincial Key Laboratory of Soil Environmental Health and Regulation, Fuzhou 350002, Fujian, China;3.Yangzhou Agriculture Environment Monitoring Station, Yangzhou 225603, Jiangsu, China;4.State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, Jiangsu, China
|
| Abstract: |
| The resolution of Digital Elevation model(DEM) plays a critical role in deciding to which level the detail of geomorphological variation in space of one region could be expressed, thus it has an essential impact on digital soil organic carbon mapping for mountainous area, and yet which is not very clear for plain area. To figure out that, digital soil organic carbon mapping for the plain area in north of Jiangsu was performed based on multiple-resolution DEMs by using Random Forest algorithm. Specifically, the relationship between soil organic carbon content and environmental factors of five type, including terrain, climate, organism, soil and parent material, was modeled to predict cultivated soil organic carbon content of plain area in north of Jiangsu at multiple resolutions, aiming to explore how DEM resolutions affect the soil organic carbon mapping for areas characterized by marked topographic variation at local scale. The results were showed as follows. (1) Degree to which influencing factors of five type affect soil organic carbon content decreased in the sequence of terrain, climate, organism, soil and parent material at different DEM resolutions. (2) While DEM resolution varied at the range of greater than 150 m, accuracy of soil organic carbon content prediction went down with increased DEM resolution, which could be primarily attributed to the fact that environmental features progressively lost details due to DEM aggregation. (3) Terrain was demonstrated to be the most dominant factor that had influence on spatial variation of soil organic carbon in the study area. In conclusion, DEM resolution between 60 and 90 meters are optimal for the soil organic carbon mapping in Subei area where is globally flat whereas varies significantly in local terrain. This research may provide theoretical basis for the attempt to increase precision of soil organic carbon mapping for the area with the similar geomorphology to this studied area. |
| Key words: soil organic carbon content Random Forest terrain DEM resolution digital soil mapping |
