土壤学报2019，Vol.56Issue(1)：44-54,11.DOI:10.11766/trxb201803070048
利用景观指数定量化评估历史土壤图制图精度
Quantitative Evaluation of Mapping Precision of Historical Soil Maps with Landscape Indices
摘要
Abstract
[Objective]It is a precondition for making proper use of these precious data like soil maps to judge mapping precision of historical soil type maps.[Method]A map marked as 1: 250 000 scale Fujian Province soil map plotted on the basis of the soil survey data of the Second National Soil Survey, was collected and digitized into a vector soil map, which was rasterized by taxonomical hierarchy.From each level of the hierarchy, 50 raster grains, varying in interval from 30 m to 8 000 m, were selected for calculation of 15 landscape indices of each grain in three taxonomical levels, i.e. great group, subgroup and family, separately, using Fragstats 4.2, a software for landscape pattern analysis, and further for analysis of responses of the landscape indices to variation of the raster grains. Five landscape indices, i.e.TA, PR, PRD, SHDI and SHEI, were selected as indices to quantitatively evaluate soil mapping precision.The landscape indices corresponding to grains 30 m×30 m in size were set as benchmarks. Comparison was made of landscape indices corresponding to grains of various sizes with the benchmarks. Given that when the absolute values of relative variability (VIV, ％) of all the selected landscape index are less than 1％, the corresponding maximum grain size is deemed as the optimal characterization grain size for data conversion from parent vector soil map to raster soil map. Then based on the functional relationship, y =-0.80×10-6 x2 +0.0 228 x + 0.0 211 (R2 = 0.9994, P ＜ 0.05) , between optimal characterization grain size and soil mapping scale, soil mapping scales could be deduced and mapping precision of the map quantitatively evaluated.[Result]Results show that landscape index is obviously grain-size dependent, and based on the trend of landscape indices varying with rising grain size, grain-size dependence of landscape indices could be divided into four types, i.e. Type I: landscape indices increasing with rising grain size, such as PAFRAC and AREA_SD; Type II: landscape indices declining with rising grain size, such as TE, LSI, COHESION and AI; Type III: landscape indices varying irregularly with rising grain size, such as LPI, DIVISION, MESH and SPLIT; and Type Ⅳ: landscape indices remaining almost unchanged or fluctuating later on with rising grain size, such as TA, PR, PRD, SHDI and SHEI. The optimal characterization grain size at the soil great group level, subgroup level and family level is 4.00 km×4.00 km, 3.45 km×3.45 km and 1.90 km×1.90 km, respectively, and their corresponding soil map scale, 1: 1 800 000, 1: 1 600 000 and 1: 850 000, respectively, which all differ significantly from the marked map scale [Conclusion]This study provides a novel and effective way and method to judge mapping precision of historical soil maps, which is of great value to correctly judge and utilize precious historical soil data.关键词
景观指数/土壤类型图/制图精度/土壤分类层次Key words
Landscape index/Soil type map/Mapping precision/Soil taxonomical level分类
农业科技引用本文复制引用
黄晶晶,于银霞,于东升,潘月,陆晓松,徐志超..利用景观指数定量化评估历史土壤图制图精度[J].土壤学报,2019,56(1):44-54,11.基金项目
国家自然科学基金项目(41571206)、国家重点研发计划专项(2016YFD0200301)和国家科技基础性工作专项(2015FY110700-S2)资助 （41571206）
Supported by the National Natural Science Foundation of China (No. 41571206) , Special Project of the National Key Research and Development Program (No. 2016YFD0200301 ) and Special Project of the National Science and Technology Basic Work (No. 2015FY110700-S2) （No. 41571206）
