Estimation of leaf area index with the Li-Cor LAI 2000 in deciduous forests
Estimation of leaf area index with the Li-Cor LAI 2000 in deciduous forests
复制标题
DOI：
10.1016/s0378-1127(97)00269-7
复制
发表时间：
1998-06-15
影响因子：
3.7
通讯作者：
Mugnozza, GS
中科院分区：
文献类型：
--
作者：
Cutini, A;Matteucci, G;Mugnozza, GS
本研究已在意大利，在主要阔叶林树种的立场。驼栎间伐林分与未间伐林分。(6.3占意大利落叶树种覆盖总面积的%），板栗，(8.5%）和欧洲山毛榉（Fagus sylvatica L.）(12.6%）。在3年（1993 - 1995）期间，使用直接（凋落物收集器）和间接方法（LAI-2000植物冠层分析仪，PCA，Li-Cor，林肯，NE，美国）收集LAI数据。通过凋落物收集估算的叶面积指数（LAI（LT），3.2 - 7.6 m（2）m（-2））在落叶林报告的数值范围内，而主成分分析法通常低估了叶面积指数（LAI（PCA），1.8 - 5.8 m（2）m（-2））。平均低估率为26.5%，与其他报道相似。间伐林分（29.1%+/-14.7%）的低估率高于未间伐林分（22.2%+/-12.2%）以及叶面积指数（LT）> 5 m（2）m（-2）的林分。相反，在LAI（LT）<5 m（2）m（-2）的林分中，PCA估计值更接近凋落物陷阱估计值（-11%+/-9.6%）。间伐林分和未间伐林分的叶面积指数（LT）分别为4.51 ± 0.92 m（2）m（-2）和5.86 ± 0.11 m（2）m（-2）。主成分分析还能够估计这一差异，得出3.14 +/-0.70 m（2）m（-2）（间伐林分）和4.47 +/-0.61 m（2）m（-2）（未间伐林分）。用这两种方法，两种林分类型之间的差异是非常显着的。虽然LAI（PCA）值总是低于LAI（LT），两个数据集之间的相关性是线性和显着的。当去掉PCA外环的阅读重新计算时，LAI（LT）和LAI（PCA）之间的相关性得到改善，LAI（PCA）的低估在12%以内。在无叶期用主成分分析法计算了林地面积指数（WAI）。该仪器能够显示间伐（0.55 +/-0.09）和未间伐（0.80 +/-0.19）之间的差异。与其他研究相似，从LAI（PCA）值中减去WAI增加了LAI（LT）的低估。两种方法估计LAT的一致性令人满意。然而，必须考虑到主成分分析方法的低估。在不同的年份和林分，低估的变化并不明显，这表明在不同的条件下，例如，比较间伐和未间伐的林分，以及测量叶面积指数的时间和空间变化时，使用主成分分析是可靠的。重叠的叶片，存在的冠层内的差距和光在地平线水平似乎是一些重要的变量，影响叶面积指数估计的主成分分析。进一步的数据校正可以大大提高PCA的性能，并产生可靠的LAI估计，即使直接参考测量的收集，强烈建议准确地知道LAI，以评估仪器性能在给定的网站。(C)1998年Elsevier Science B.V.
The present study has been performed in Italy, in stands of the main broad-leaved forest species. Thinned and unthinned stands of Quercus cerris L. (6.3% of total area covered by deciduous species in Italy), Castanea sativa L., (8.5%) and Fagus sylvatica L. (12.6%) were selected from 15 permanent plots. LAI data have been collected during the 3 years (1993-1995), using both direct (littertraps) and indirect methods (LAI-2000 Plant Canopy Analyzer, PCA, Li-Cor, Lincoln, NE, USA). LAI estimation by litter collection (LAI(LT), 3.2-7.6 m(2) m(-2)) was in the range of values reported for deciduous forest, while the PCA method generally underestimated the LAI (LAI(PCA), 1.8-5.8 m(2) m(-2)). Average underestimation was 26.5%, being similar to other reports. The underestimation was higher in thinned (29.1% +/- 14.7%) than in unthinned (22.2% +/- 12.2%) stands and in stands characterised by a LAI(LT) > 5 m(2) m(-2). On the contrary, in stands with LAI(LT) < 5 m(2) m(-2), PCA estimates were closer to littertraps ones (-11% +/- 9.6%). On the average, LAI(LT) was 4.51 +/- 0.92 m(2) m(-2) and 5.86 +/- 0.11 m(2) m(-2), for thinned and unthinned stands, respectively. Also PCA was able to estimate this difference, arriving at 3.14 +/- 0.70 m(2) m(-2) (thinned stands) and 4.47 +/- 0.61 m(2) m(-2) (unthinned stands). With both methods, the difference between the two stand types was strongly significant. Although LAI(PCA) values were always below LAI(LT), the correlation between the two data sets was linear and significant. When recalculated omitting the reading of the external PCA ring, the correlation between LAI(LT) and LAI(PCA) improved, and the underestimation of LAI(PCA) was within 12%. Woody area index (WAI) was evaluated with the PCA during the leafless period. The instrument was able to show the difference between thinned (0.55 +/- 0.09) and unthinned stands (0.80 +/- 0.19). Similar to other studies, the subtraction of WAI from LAI(PCA) values increased the underestimation of LAI(LT). The agreement between the two methods for LAT estimation was satisfactory. Nevertheless, the underestimation by the PCA method must be taken into account. Over different years and stands, the variability of underestimation was not marked, pointing to a reliable use of PCA in different conditions, as, for example, the comparison of thinned and unthinned stands and in measuring temporal and spatial variations of LAI. The overlapping of leaves, the presence of gaps within the canopy and light at the horizon level seem to be some of the important variables that influence LAI estimation by the PCA. Further corrections of the data can improve substantially the performance of the PCA and produce reliable LAI estimates even though the collection of direct reference measurements is strongly recommended to know exactly LAI, in order to assess instrument performance in a given site. (C) 1998 Elsevier Science B.V.