基于不同尺度样地的森林面积估算对比分析1）--以鞍山市为例

摘要

研究一类调查与二类调查的差异，分析现有森林资源样地调查方式的效率和误差，对比分析调整样地数量和样地面积大小后对市级区域森林面积估算的精度。以鞍山市为研究区，以二类调查小班数据为基础，以现行4 km×8 km样地间隔的标准样地数量为最大样本量，分析了在样方边长分别为25、30、250 m，1、2、3、4 km与样地数量分别为现行数量的1／2、1／3、1／4和1／5等情况下样地估算森林面积总体的精度。结果表明：样地面积越大，像元的混合度越大；现行抽样方式精度较高，误差仅为0．65％；当样地数量为现有样地数量的1／2，样地大小为2 km×2 km时，是较为准确估算鞍山市森林面积值的最佳抽样方式，误差为02．9％；虽然采样率为1／3，边长为3 km和4 km方式的精度更高，但是调查的面积更大。%We studied the differences between two types of surveys, efficiency and error analysis of an existing sample-plot sur-vey method, and the accuracy of the comparative analysis of municipal estimation after adjusting the number and area of sample-plots.We took Anshan as the study area, subcompartment data as reference, and the current 4 km×8 km interval standard plot-samples as the largest sample size.Then, we changed the quadrats side length from 25 m, 30 m, 250 m, 1 km, 2 km, 3 km to 4 km and the number of sample-plots from 1/2, 1/3, 1/4 to 1/5 of the current number. Finally, we estimated the accuracy of the overall.The larger of the plot area was, the greater of the pixel mixedness was.The accuracy of existing sampling was higher with the error of 0.65%. When the number of sample-plots reached the half of the existing sample number, the size of the plot was 2 km×2 km, the estimation of the forest area ratio in Anshan was best, and the er-ror was 0.29%.When the sampling rate was 1/3, the side lengths were 3 km and 4 km with higher accuracy, but with lar-ger investigation area.