Optimizing the sampling scheme for LAI-2000 measurements in a boreal forest
Optimizing the sampling scheme for LAI-2000 measurements in a boreal forest
复制标题
DOI：
10.1016/j.agrformet.2011.10.002
复制
发表时间：
2012-03
影响因子：
6.2
通讯作者：
Titta Majasalmi;M. Rautiainen;P. Stenberg;H. Rita
中科院分区：
文献类型：
--
作者：
Titta Majasalmi;M. Rautiainen;P. Stenberg;H. Rita
叶面积指数（LAI）的实地测量是许多生态学和遥感应用所必需的，但这是一项费时费力的工作。光学叶面积指数测量的采样设计主要受传感器的视场以及卫星图像的放大过程和空间分辨率的影响。一个优化的采样设计将提高叶面积指数测量的准确性和减少所需的现场工作的时间。然而，尚未对北方森林叶面积指数测量的抽样设计进行系统比较。本研究的目的是分析抽样设计对LAI-2000植物冠层分析仪在北方森林中获得的LAI估计的准确性的影响。田间数据收集自芬兰芬兰中部Hyytiälä的6个研究中心。樟子松、欧洲云杉和垂枝桦各占2个优势样地。每个站点都有一个规则的9 × 9网格的LAI测量点。使用定期和系统的抽样技术对所有地点进行了研究。统计分析用于检查随机抽样中达到给定精度所需的特定地点的测量点数量，并作为定期和系统抽样的起点。结果表明，12个随机定位的测量点是足够的特征与标准误差小于0.15LAI单位的站点特定的LAI。如果系统地组织12个采样点，则平均标准误差减小到0.06个单位，变异系数小于0.03。
Field measurements of leaf area index (LAI) are needed for many ecological and remote sensing applications, but they are laborious and time-consuming. The sampling designs of optical LAI measurements have been mainly driven by the field-of-view of the sensor as well as the upscaling process and spatial resolution of satellite imagery. An optimized sampling design would improve the accuracy of the LAI measurements and reduce the time needed for field work. However, a systematic comparison of sampling designs for LAI measurements in boreal forests has not been done. The aim of this study was to analyze the effect of sampling designs on the accuracy of LAI estimates obtained with the LAI-2000 Plant Canopy Analyzer in a boreal forest. Field data was collected from six study sites in Hyytiälä, central Finland. Two sites were dominated by Pinus sylvestris, two sites by Picea abies, and two sites by Betula pendula. Each site had a regular 9×9 -grid of LAI measurement points. All sites were studied using regular and systematic sampling techniques. Statistical analysis was used to examine the site-specific number of measurement points needed in random sampling to achieve a given precision, and as a starting point for regular and systematic sampling. Results indicated that twelve randomly located measurement points are enough to characterize site-specific LAI with standard error less than 0.15LAI units. If the twelve sampling points are organized systematically, the average standard error is reduced to 0.06 units and the coefficient of variation to less than 0.03.