摘要
我国宏观森林面积是通过森林资源连续清查获得的,该方法采用的是系统抽样方法。从理论上讲,一个样地的森林面积计量可以有两类方法:0-1离散变量法和0-1之间的连续比例变量法。我国目前采用的是前一类,即一个样地或是森林,或不是森林,没有其他状态。在确定一个样地是否是森林时,又有两种方法可选择:一是点定法,即根据样地的某个点(如样地西南角)的地类确定;二是优势地类法,即根据样地内森林比例是否占优势确定。这样对样地森林面积的计量总共可以有3种不同的估计方法。采用计算机模拟的方法对这3种方法进行试验研究,选择块状(连片不规则分布)和带状两种空间分布形式的森林作为研究对象,这两种分布分别代表山区(林区)和平原农区的森林分布状况。研究结果表明:1)连续变量法效率最高,且无偏;2)点定法也无偏,但效率较连续变量法低;3)优势地类法则是有偏的,尤其对于带状森林,偏差可能很大,偏差大小决定于林带的宽度及样地大小。以上结论对于目前正在探讨与实践的遥感大样地来说也有参考价值。对于遥感大样地来说,应该选择无偏且有最高估计效率的连续变量法。
National or regional forest area in China is usually estimated by the Continuous Forest Inventory ( CFI) ,which is conducted in a systematic sampling manner. In theory,two kinds of methods can be used to determine the forest area of a plot:0 -1 discrete variable and continuous proportion variable ( CV) between 0 -1 . The former is used at present in China,that is,a plot is either forest or not forest,and no other condition. There are two methods for 0 - 1 discrete variable determination, one is point land based (PLB) ,for example,the southwest comer of a plot,the other is dominant land based(DLB) . Then,there are total three methods to determine the forest area measurement of a plot. In this paper,computer simulation is used for experimental research to these three methods and the irregular area and strip distributed forest were selected as study objects, these two types of distribution are for mountain ( forest region) and plain agricultural region respectively. The result shows that:l. CV has the highest sampling efficiency,and it is unbiased;2. PLB is also unbiased,but its efficiency is lower than CV. 3. DLB is seriously biased,especially for the strip forest, and gets larger when plot size gets larger. The result might be considered in remote sensing large-plot-sampling,which is deployed in China at present,and CV method should be chosen for this kind of sampling.
作者
孟源源
葛宏立
MENG Yuanyuan;GE Hongli(School of Environmental & Resource Sciences, Zhejiang A & F University;Zhejiang Provincial Key Laboratory of Carbon Cycling in Forest Ecosystems and Carbon Sequestration,Lin‘an 311300 , Zhejiang, China)
出处
《林业资源管理》
北大核心
2016年第2期49-55,共7页
Forest Resources Management
基金
国家科技支撑计划团队任务"低效公益林更新改造和健康维持技术研究与示范"(2012BAD22B0503)
