基于光学仪器法测定谷地云冷杉林叶面积指数的季节变化

利用半球摄影法(DHP)和LAI-2000植物冠层分析仪两种光学仪器法(间接法)以及凋落物法(直接法),研究了小兴安岭谷地云冷杉林叶面积指数(LAI)的季节变化,并构建了不同季节直接法与间接法测定的LAI间的相关关系.结果表明: 在整个试验期间,DHP测定的LAI比直接法测定值低估40%～48%, LAI-2000植物冠层分析仪的低估范围为15%～26%;不同时期直接法与DHP和LAI-2000植物冠层分析仪测定的LAI均显著相关, 且均可合并为A、B、C 3类预测模型, 可以分别预测5和11月, 6、9和10月, 7和8月的LAI.本研究结果可为高效、准确地测定针叶林LAI的季节变化提供参考.     

利用直接法和间接法测定针阔混交林叶面积指数的季节动态

利用光学仪器法能够快速、高效地测定森林生态系统的叶面积指数(leaf area index, LAI)。然而, 评估该方法测定针阔混交林LAI季节动态准确性的研究较少。该研究基于凋落物法测定了小兴安岭地区阔叶红松(Pinus koraiensis)林LAI的季节动态, 其结果可代表真实的LAI。参考真实的LAI, 对半球摄影法(digital hemispherical photography, DHP)和LAI-2000植物冠层分析仪测定的有效叶面积指数(effective LAI, L_e)进行了评估。首先对DHP测定LAI过程中采用的不合理曝光模式(自动曝光)进行了系统校正。同时, 测定了光学仪器法估测LAI的主要影响因素(包括木质比例(woody-to-total area ratio, α)、集聚指数(clumping index, Ω_E)和针簇比(needle-to-shoot area ratio, γ_E))的季节变化。结果表明: 3种不同方法测定的LAI均表现为单峰型的季节变化, 8月初达到峰值。从5月至11月, DHP测定的L_e比真实的LAI低估50%-59%, 平均低估55%; 而LAI-2000植物冠层分析仪测定的L_e比真实的LAI低估19%-35%, 平均低估27%。DHP测定的L_e 经过自动曝光, α、Ω_E和γ_E校正后, 精度明显提高, 但仍比真实的LAI低估6%-15%, 平均低估9%; 相对而言, LAI-2000植物冠层分析仪测定的L_e经过α、Ω_E和γ_E校正后, 精度明显提高, 各时期与真实的LAI的差异均小于9%。研究结果表明, 考虑木质部和集聚效应对光学仪器法的影响后, DHP和LAI-2000植物冠层分析仪均能相对准确地测定针阔混交林LAI的季节动态, 其中, DHP的测定精度高于85%, 而LAI-2000植物冠层分析仪的测定精度高于91%。     

Measuring seasonal dynamics of leaf area index in a mixed conifer-broadleaved forest with direct and indirect methods

利用光学仪器法能够快速、高效地测定森林生态系统的叶面积指数(leaf area index, LAI)。然而, 评估该方法测定针阔混交林LAI季节动态准确性的研究较少。该研究基于凋落物法测定了小兴安岭地区阔叶红松(Pinus koraiensis)林LAI的季节动态, 其结果可代表真实的LAI。参考真实的LAI, 对半球摄影法(digital hemispherical photography, DHP)和LAI-2000植物冠层分析仪测定的有效叶面积指数(effective LAI, L_e)进行了评估。首先对DHP测定LAI过程中采用的不合理曝光模式(自动曝光)进行了系统校正。同时, 测定了光学仪器法估测LAI的主要影响因素(包括木质比例(woody-to-total area ratio, α)、集聚指数(clumping index, Ω_E)和针簇比(needle-to-shoot area ratio, γ_E))的季节变化。结果表明: 3种不同方法测定的LAI均表现为单峰型的季节变化, 8月初达到峰值。从5月至11月, DHP测定的L_e比真实的LAI低估50%-59%, 平均低估55%; 而LAI-2000植物冠层分析仪测定的L_e比真实的LAI低估19%-35%, 平均低估27%。DHP测定的L_e 经过自动曝光, α、Ω_E和γ_E校正后, 精度明显提高, 但仍比真实的LAI低估6%-15%, 平均低估9%; 相对而言, LAI-2000植物冠层分析仪测定的L_e经过α、Ω_E和γ_E校正后, 精度明显提高, 各时期与真实的LAI的差异均小于9%。研究结果表明, 考虑木质部和集聚效应对光学仪器法的影响后, DHP和LAI-2000植物冠层分析仪均能相对准确地测定针阔混交林LAI的季节动态, 其中, DHP的测定精度高于85%, 而LAI-2000植物冠层分析仪的测定精度高于91%。     

自动曝光对半球摄影法测量叶面积指数及其季节变化的影响

自动曝光是影响半球摄影法(DHP)测量叶面积指数(LAI)精度的重要误差源之一.本研究基于小兴安岭地区的阔叶红松林、白桦次生林、红松人工林和兴安落叶松人工林,利用DHP和LAI-2200植物冠层分析仪分别测量6—9月每个月中旬的LAI,首先比较两种方法测量LAI的差异性,再检验森林类型和测量时期对建立两种方法测定值间的相关关系是否存在显著影响,最后构建适于校正不同森林类型不同时期自动曝光对DHP测量LAI产生误差的经验模型.结果表明: 4种森林类型4个时期内,在自动曝光设置下DHP测量的LAI比LAI-2200测量值低估20%～49%;森林类型对构建两种方法测量LAI值的经验模型不存在显著影响,而测量时期存在显著影响.本研究构建的A、B两种分类经验模型,分别适用于校正4种森林类型在6和9月、7和8月DHP测量的LAI.经分类经验模型校正后,DHP测量4种森林类型4个时期的LAI值提高了45%～79%,测量精度可提高到83%～94%.通过DHP和LAI-2200测量LAI值间的经验模型,可有效校正自动曝光对DHP测量LAI的影响,极大地提高其测量精度,为使用DHP快捷、高效地测量不同森林类型的LAI及其季节动态提供技术支持.     

广西桉树人工林叶面积指数的估测及其校正

叶面积指数(Leaf area index, LAI)是森林生态系统重要的结构参数,通过遥感技术可反演区域LAI,但其可靠性需要地面准确的实测数据进行验证。选取广西国有高峰林场不同林龄的桉树(Eucalyptus robusta)人工林为对象,以异速生长法(Allometry)为对照,综合利用植物冠层分析仪法(LAI-2200)、跟踪辐射和冠层结构分析仪法(TRAC)、半球摄影法(DHP)以及地基激光雷达法(TLS)等间接法估测样地的LAI,并考虑木质成分以及聚集效应影响,进行相应的校正处理,为地面快速、准确测量桉树人工林LAI提供参考。结果表明：桉树人工林的比叶面积为125.37±13.38 cm~2/g,通过Allometry获得的LAI变化范围在1.65—3.84,平均为2.73,不同林龄间的差异均显著(P<0.05),随着林龄的增加呈现先增加后减少的趋势。在未校正情况下,LAI-2200、TRAC、DHP、TLS估算的LAI存在显著差异(P<0.05)。与对照相比,LAI-2200在幼龄林和过熟林中估算误差最小,TRAC在成熟林中估算误差最小。相对于完全去除法,利用...     

小兴安岭白桦次生林叶面积指数的估测

叶面积指数(LAI)是量化冠层结构最常用的参数之一,准确估测LAI对森林生态系统结构特性的研究具有重要意义.利用半球摄影图像法和LAI-2000法及半球摄影图像法结合凋落物法估测了小兴安岭白桦次生林LAI及其动态变化.首先对该林型叶凋落末期(11月初)的半球摄影图像进行合理校正(包括木质部分所占比例α,冠层水平集聚指数ΩE,校正值作为该时期常绿树种的真实LAI(LAIt),结合各调查期的凋落物数据,得到落叶季节(7-11月)的LAIt,并以该值为参考值,对比分析了两种光学仪器法估测值.结果表明:两种光学仪器法在LAI最大时期低估(分别低估2.83％、6.20％),其他时期显著高估(平均高估118.13％、89.34％),但两种光学仪器法与探讨方法估测值存在很好的相关性:LAIt=-1.1393+1.0934·LAIHP,R2=0.80; LAIt=-0.1712+0.6259·LAILAI-2000,R2=0.83.研究结果可为将来方便、快捷、准确的估测白桦次生林的LAI提供参考.     

基于数码相片Gamma校正的水稻叶面积指数估算

随着数码相机的日益普及,利用数码相机进行作物叶面积指数(LAI)测量不断得以应用。由于数码相机成像时会对入射光辐射强度进行Gamma编码变换,输出的相片DN(Digital Number)值与入射光辐射强度呈非线性关系,会造成在确定相片中植被叶片与背景的分割阈值时出现误差,并最终导致LAI估算存在较大不确定性。以水稻为研究对象,获取不同生长期水稻冠层相片并结合同步LAI 2000测量的LAI数据,基于相片Gamma校正原理,对水稻不同生长期冠层相片进行Gamma校正,在此基础上利用冠层孔隙率方法,估算不同生长期水稻LAI。结果表明,经过Gamma校正相片估算的水稻LAI总体精度有显著提高,相片估算的IMAGE LAI与LAI-2000测量值比较的决定系数达到0.71(P0.05)。在整个观测期内,两种方法观测的LAI值在时间变化趋势上表现一致,但在不同生长期内存在差别,在水稻分蘖期和拔节期相片估算的IMAGE LAI要高于LAI-2000测量值,孕穗期到抽穗期期间IMAGE LAI低于LAI-2000测量值,乳熟期到成熟期IMAGE LAI又高于LAI-2000的观测结果。     

帽儿山地区森林冠层叶面积指数的地面观测与遥感反演

叶面积指数(leaf area index,LAI)是陆地生态系统最重要的结构参数之一,遥感和基于冠层孔隙率模型的光学仪器观测是快速获取LAI的有效方法,但由于植被叶片的聚集效应,这些方法通常只能获取有效叶面积指数(effective LAI,LAIe).本文以东北林业大学帽儿山实验林场为研究区,利用LAI2000观测森林冠层LAIe,并结合TRAC观测的叶片聚集度系数估算了森林冠层LAI,并通过分析基于Landsat5-TM数据计算的不同植被指数与LAIe之间的关系,建立了该区森林LAI遥感估算模型.结果表明:研究区阔叶林的LAI和LAIe基本相当,而针叶林的LAI比LAIe大27%;减化比值植被指数(reduced simple ratio,RSR)与该区LAIe的相关性最好(R2=0.763,n=23),最适合该区LAI的遥感提取.当海拔<400 m时,LAI随海拔高度的上升而快速增大;当海拔在400～750 m时,LAI随海拔高度的上升缓慢增大;当海拔>750 m时,LAI呈下降趋势.研究区森林冠层LAI与森林地上生物量存在显著的正相关关系.     

利用3种校正方案提高间接法测定兴安落叶松人工林叶面积指数的精度

木质部和集聚效应是影响间接法测定叶面积指数(LAI)精度的主要因素, 尤其是木质部的校正一直存在争议。针对这一问题, 该研究首先利用半球摄影法(DHP)和LAI-2000植物冠层分析仪法(LAI-2000法) 2种间接法测定了小兴安岭兴安落叶松(Larix gmelinii)人工林叶面积最大时期的有效LAI (L_e), 然后提出了A、B、C 3种校正方案来提高间接法的测定精度。同时, 利用凋落物法和异速生长方程法2种直接法测定LAI, 以凋落物法测定值为标准来评估3种校正方案的校正效果, 并检验天顶角范围对校正结果是否存在显著影响。结果表明: 在0-45° (1-3环)、0-60° (1-4环)、45°-60° (4环)及0-75° (1-5环) 4种不同天顶角范围内, DHP测定的L_e比凋落物法、异速生长方程法测定值分别低估19%-32%和8%-29%; 而LAI-2000法也得到相似的结论, 分别低估9%-30%和8%-28%。虽然校正方案A高估了木质部对LAI的贡献, 但在45º-60º天顶角范围内, 能有效地校正DHP测定的L_e, 在1-3环和1-4环天顶角范围内, 能有效地校正LAI-2000法测定的L_e。4种天顶角范围内, 校正方案B均能有效地校正DHP测定的L_e。整体来看, 4种天顶角范围内, 校正方案C对DHP和LAI-2000法测定值的校正效果均优于其他2种方案。研究结果表明除木质部和集聚效应外, 天顶角范围的选择也是决定间接法测定LAI精度的重要因素。     

小麦LAI-2000观测值对辐亮度变化的响应

利用LAI-2000植物冠层分析仪和ASD光谱仪,通过固定点拔节期冬小麦叶面积指数(LAI)观测实验和同步光谱辐亮度实验,测量了晴天条件下,固定点冬小麦从中午至傍晚24个不同时刻的LAI值及对应的辐亮度。继之分析了此段时间内测得的定点冬小麦LAI值分别与对应时刻可见光和近红外谱段的天空光辐亮度、总辐亮度和太阳直射辐亮度值之间的相关性。结果表明,无论波长小于490nm的谱段,还是波长大于490 nm的谱段,LAI与天空光辐亮度、总辐亮度和太阳直射辐亮度都呈负相关,相关系数(R2)高达0.8左右;尤其LAI与天空光辐亮度的负相关性最高,这种相关性随着波长的增大而减小。LAI与各谱段天空光辐亮度的相关性特征可为LAI-2000晴天观测的LAI值归一化修正处理提供一种新思路和技术途径,以消除太阳直射光的影响,从而解决LAI-2000只能在晴天观测的局限性,拓展LAI-2000在晴天观测条件下的适用性。     

新疆喀纳斯国家自然保护区植被叶面积指数观测与遥感估算

叶面积指数(Leaf Area Index,LAI)是重要的植被结构参数,调控着植被与大气之间的物质与能量交换,在生态环境脆弱的我国西北部开展植被LAI的研究对阐明该地区植被对气候变化和人类活动的响应特征具有重要的科学意义.利用LAI-2200和TRAC仪器观测了新疆喀纳斯国家级自然保护区森林和草地的有效叶面积指数(LAIe)和真实LAI,构建了其遥感估算模型,生成了研究区LAIe和LAI的空间分布图.在此基础上,分析了LAI随地形因子(海拔、坡度、坡向)的变化特征,探讨了将其应用于估算研究区森林生物量密度的可行性,并评估了研究区MODIS LAI产品的精度.结果表明:研究区阔叶林、针阔混交林、针叶林、草地LAIe的平均值分别为4.40、3.18、2.57、1.76,LAI的平均值分别为4.76、3.93、3.27、2.30.LAIe和LAI的高值主要集中分布在湖泊和河流附近;植被LAI随海拔、坡度和坡向的变化表现出明显的垂直地带性的特点.LAI随海拔和坡度的增加呈现先增加后减小的变化趋势,坡向对针叶林和草地LAI的影响明显,但对阔叶林和针阔混交林LAI的影响较弱;森林生物量密度(BD)随LAI增加而线性增加(BD=44.396LAI-25.946,R2=0.83),研究区森林生物量密度平均值为120.3 t/hm2,估算的总生物量为5.0×l06 t;MODIS LAI产品与利用TM数据生成的LAI之间具有一定的相似性(森林R2=0.42,草地R2=0.53),但森林和草地的MODIS LAI产品分别比利用TM数据生成的LAI偏低16.5％和24.4％.     

Test of Accuracy of LAI Estimation by LAI-2000 under Artificially Changed Leaf to Wood Area Proportions

The accuracy of LAI-2000 Plant Canopy Analyzer for leaf (LAI) and plant (PAI) area indexes measurements was tested in 20-year-old Norway spruce stand using the reduction of canopy biomass. Needle and branch areas were reduced progressively upward every one meter. Values of effective leaf area index (LAI_e), as an uncorrected product of LAI-2000, were compared with directly estimated LAI and PAI values after each reduction step. LAI-2000 underestimates PAI and LAI values according to LAI-2000 rings readings, and varied proportions between leaf and wood areas. The values of LAI_c have been increased with decreasing of the view angle of the relevant LAI-2000 rings. Therefore, the underestimation of LAI becomes smaller when the readings near the horizon are masked. More accurate results, for projected LAI (LAI_p) calculation, are produced by LAI-2000 when some dense grids of measurement points and the most vertical ring readings (0 –13 °) are used. Correction factor 1.6 is possible to use for unreduced canopy hemi-surface LAI estimation, when the last rings (i.e. 5^th and 4^th rings, 47 –74 °) are excluded. Correction factor of 1.25 can be used to compute LAI_p if the angle readings under 43 °are also masked.     

Improving accuracy of optical methods in estimating leaf area index through empirical regression models in multiple forest types

Key message

We developed the empirical regression models relating the direct LAI and optical LAI from initial leaf out to the leaf fall in different forest types in China.

Abstract

Optical methods have usually been used to estimate the leaf area index (LAI) in a forest stand because of rapidity and reduced labor requirements. However, few studies have reportedly improved the accuracy of the optical LAI estimates for seasonal dynamics using empirical models in different forest types. In the present study, we directly measured the seasonal dynamics of LAI from leaf out to leaf fall based on litter collection (defined as direct LAI) in a mixed evergreen–deciduous forest, an evergreen forest and a deciduous forest. Meanwhile, the effective LAI was estimated using digital hemispherical photography (DHP) and LAI-2000 instruments. Our main objective was to explore the seasonal changes in the relationship between direct LAI and effective LAI values and to find the best LAI empirical estimation model in different forest types. The season-dependent models relating direct LAI and effective LAI in each period were developed through a power function regression model in several forest types. Then, significance tests were applied to compare the different season-dependent models. The analysis showed that the season-dependent models can be merged into different aggregated models depending on forest types and optical methods. We confirm that the seasonal changes in LAI in different forest types can be fully estimated through aggregated models using both DHP and LAI-2000 methods with accuracies of more than 87 and 92 %, respectively. Meanwhile, our results suggest that the forest type (i.e., species composition of forest stand) and optical method should be seriously considered to correctly and quickly estimate the seasonal changes of LAI through the aggregated models.

     

Estimation of deciduous forest leaf area index using direct and indirect methods

This study evaluated one semi-direct and three indirect methods for estimating leaf area index (LAI) by comparing these estimates with direct estimates derived from litter collection. The semi-direct method uses a thin metallic needle to count a number of contacts across fresh litter layers. One indirect method is based on the penetration of diffuse global radiation measured over the course of a day. The second indirect method uses the LAI-2000 plant canopy analyser (PCA) which measures diffuse light penetration from five different sky sectors simultaneously. The third indirect method uses the Demon portable light sensor to measure the penetration of direct beam sunlight at different zenith angles over the course of half a day. The Poisson model of gap frequency was applied to estimate plant area index (PAI) from observed transmittances using the second and third methods. Litter collection from 11 temperate decidous forests gave values of LAI ranging from 1.7 to 7.5. Estimates based on the needle method showed a significant linear relationship with LAI values obtained from litter collections but were systematically lower (by 6–37%). PAI estimates using all three indirect techniques (fixed light sensor system, LAI-2000 and Demon) showed a strong linear relationship with LAI derived from litter collection. Differences, averaged over all forest stands, between PAI estimates from each of the three indirect methods and LAI from litter collections were below 2%. If we consider that LAI=PAI–WAI (wood area index) then, all three indirect methods underestimated LAI by an additional factor close to the value of WAI. One reason could be a local clumping of architectural canopy components: in particular, the spatial dispositions of branchlets and leaves are not independent, leading to a non-random relationship between the distributions of these two canopy components.     

Improving the accuracy of indirect methods in estimating leaf area index using three correction schemes in a Larix gmelinii plantation

Aims Woody materials and clumping effects are key error sources in estimating leaf area index (LAI) by optical methods. However, how to correct the error caused by woody materials has not reached consensus. The aims of this study are (1) to evaluate the accuracy of optical methods for estimating effective LAI (L_e) in a deciduous needle leaf forest stand, and (2) to develop a practical correction scheme to improve the accuracy of optical methods in estimating LAI.Methods L_ewas estimated by two indirect methods (i.e., digital hemispherical photography (DHP) and LAI-2000 plant canopy analyzer method (LAI-2000 method) in an annual maximum leaf area period in a Larix gmelinii plantation. Then, we developed three correction schemes to improve the accuracy of indirect methods in estimating LAI. Meanwhile, two direct methods (i.e., litter collection and allometry methods) were used to estimate LAI. Taking LAI from litter collection as a reference, we evaluated the effectiveness of three correction schemes and tested the influence of zenith angle ranges on the correction results.Important findings With zenith angle ranges of 0-45° (rings 1-3), 0-60° (rings 1-4), 45°-60° (ring 4) and 0-75° (rings 1-5), L_eobtained from DHP underestimated LAI from both litter collection and allometry by 19%-32% and 18%-29%, respectively. L_efrom LAI-2000 method with four zenith angles also underestimated LAI from both litter collection and allometry by 9%-30% and 8%-28%, respectively. Although the contribution of woody materials to LAI was overestimated in correction scheme A, it was effective in correcting L_efrom DHP with zenith angles of 45º-60º (ring 4), and also effective for L_efrom LAI-2000 method with zenith angles of rings both 1-3 and 1-4. Correction scheme B was all effective in correcting L_efrom DHP with four zenith angle ranges. Generally, correction scheme C was more effective than other two schemes in correcting L_efrom both DHP and LAI-2000 method with four zenith angle ranges. These results indicate that the zenith angle range is a key factor for determining the accuracy of optical methods in estimating LAI besides woody materials and clumping effects.