玛纳斯县黏土实测数据与历史数据的比较差异性分析
2018-08-14魏影武红旗张文太轩俊伟
魏影 武红旗 张文太 轩俊伟
摘要:以新疆玛纳斯县为研究区,以54个土壤调查样点实测数据为基准,对比分析了玛纳斯县1∶5万新疆土壤数据库(XSD)、1∶100万面向陆面模拟的中国土壤数据集(LSM)和1∶100万世界土壤数据库(HWSD)土壤黏粒含量的差异。结果表明,对于点数据而言,XSD平均绝对误差和平均相对误差均最大,分别为133.1 g/kg和45.87%。三种数据库的土壤黏粒数据与实测数据均有不同程度的差异。玛纳斯县1∶5万土壤质地图显示玛纳斯县有黏土分布,黏土分布面积为141.8 km2。从玛纳斯县土壤黏粒空间分布图可看出,同一区域不同数据库之间存在差异,HWSD高于300 g/kg土壤黏粒分布面积最大,为1 621.7 km2,而XSD高于 300 g/kg土壤黏粒的面积仅为104.3 km2,LSM不存在高于300 g/kg的土壤黏粒分布区。通过实测值检验了三种数据库之间的差异,表明在县域尺度XSD空间数据与实测值更接近,比LSM数据和HWSD数据更为可信。
关键词:土壤黏粒;空间分布;不确定性; ArcGIS;玛纳斯县;土壤数据库
中图分类号:S155+.3(245)文献标识号:A文章编号:1001-4942(2018)05-0092-06
Abstract Taking Manas County as research area and based on the measured data of 54 soil survey samples, we compared and analyzed the difference of soil clay content between 1∶50 000 Xinjiang Soil Database of Manas County (XSD), 1∶1 000 000 China Dataset of Soil Properties for Land Surface Modeling (LSM) and 1∶1 000 000 Harmonized World Soil Database (HWSD). For the point data, the average absolute error and the average relative error of XSD were the largest, reaching 133.1 g/kg and 45.87%, respectively. All soil clay data of the three databases are quite different from those of the actual measurement results. The 1∶50 000 soil texture map of Manas County showed that it has clay soils, and the area was 141.8 km2. From the spatial distribution of soil clay in Manas County, we can see that there are differences between different databases in the same area. The area with soil clay content over 300 g/kg was 1 621.7 km2 in the HWSD and 104.3 km2 in the XSD, but there was no clay soil area for LSM. Through comparing the differences among the three databases with the actual measuring, we found that XSD data was closer to the actual measured ones and was more reliable than LSM and HWSD data at county scale.
Keywords Soil clay; Spatial distribution; Uncertainty; ArcGIS; Manas County; Soil database
中國土壤数据库工作开始于20世纪80年代[1]。1979年新疆开展了第二次土壤调查工作[2],并建立了数字化1∶100万新疆土壤类型图[3]。新疆地域广阔,由于地形、地貌等因素限制,土壤调查范围有限,历史土壤数据可能存在较大的不确定性。1∶5万土壤类型图,虽然精度高,但不易获取土壤黏粒数据,1∶100万土壤黏粒数据虽然方便获取,但精度较低,获取精确的土壤数据尤为重要。研究者对样点数量不同[4]、研究方法不同[5-7]、补充样本[8]等空间不确定性已做了大量研究,Budiman等利用历史数据和实测数据研究了土壤碳的变化,采用数字土壤制图技术对不同地方进行不确定性验证[9]。……
