Direct and indirect estimates of Leaf Area Index (LAI) for lodgepole and loblolly pine stands

We compared direct and indirect estimates of leaf area index (LAI) for lodgepole and loblolly pine stands. Indirect estimates of LAI using radiative methods of the LI-COR LAI-2000 Plant Canopy Analyzer (PCA) did not correlate with allometric estimates for lodgepole pine, and correlated only weakly with litter-trap estimates for loblolly pine. The PCA consistently under-estimated LAI in lodgepole pine stands with high LAI, and over-estimated LAI in the loblolly pine stands with low LAI. We developed a physical model to test the hypothesis that the PCA may under-estimate LAI in high leaf area stands because of increased foliage overlap and, therefore, increased selfshading. Radiative estimates of LAI using the PCA for the physical model were consistenly lower than allometric measures. Results from the physical model suggested that increased foliage overlap decreased the ability of the PCA to accurately estimate LAI. The relationship between allometric and radiative measures suggested an upper asymptote in LAI estimated using the PCA. The PCA may not accurately estimate LAI in stands of low or high leaf area index, and the bias or error associated with these estimates probably depends on species and canopy structure. A species specific correction factor will not necessarily correct bias in LAI estimates using the PCA.     

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.     

Estimation of Leaf Area Index and Plant Area Index of a Submerged Macrophyte Canopy Using Digital Photography

Non-destructive estimation using digital cameras is a common approach for estimating leaf area index (LAI) of terrestrial vegetation. However, no attempt has been made so far to develop non-destructive approaches to LAI estimation for aquatic vegetation. Using the submerged plant species Potamogeton malainus, the objective of this study was to determine whether the gap fraction derived from vertical photographs could be used to estimate LAI of aquatic vegetation. Our results suggested that upward-oriented photographs taken from beneath the water surface were more suitable for distinguishing vegetation from other objects than were downward-oriented photographs taken from above the water surface. Exposure settings had a substantial influence on the identification of vegetation in upward-oriented photographs. Automatic exposure performed nearly as well as the optimal trial exposure, making it a good choice for operational convenience. Similar to terrestrial vegetation, our results suggested that photographs taken for the purpose of distinguishing gap fraction in aquatic vegetation should be taken under diffuse light conditions. Significant logarithmic relationships were observed between the vertical gap fraction derived from upward-oriented photographs and plant area index (PAI) and LAI derived from destructive harvesting. The model we developed to depict the relationship between PAI and gap fraction was similar to the modified theoretical Poisson model, with coefficients of 1.82 and 1.90 for our model and the theoretical model, respectively. This suggests that vertical upward-oriented photographs taken from below the water surface are a feasible alternative to destructive harvesting for estimating PAI and LAI for the submerged aquatic plant Potamogeton malainus.     

Standardizing the Protocol for Hemispherical Photographs: Accuracy Assessment of Binarization Algorithms

Hemispherical photography is a well-established method to optically assess ecological parameters related to plant canopies; e.g. ground-level light regimes and the distribution of foliage within the crown space. Interpreting hemispherical photographs involves classifying pixels as either sky or vegetation. A wide range of automatic thresholding or binarization algorithms exists to classify the photographs. The variety in methodology hampers ability to compare results across studies. To identify an optimal threshold selection method, this study assessed the accuracy of seven binarization methods implemented in software currently available for the processing of hemispherical photographs. Therefore, binarizations obtained by the algorithms were compared to reference data generated through a manual binarization of a stratified random selection of pixels. This approach was adopted from the accuracy assessment of map classifications known from remote sensing studies. Percentage correct () and kappa-statistics () were calculated. The accuracy of the algorithms was assessed for photographs taken with automatic exposure settings (auto-exposure) and photographs taken with settings which avoid overexposure (histogram-exposure). In addition, gap fraction values derived from hemispherical photographs were compared with estimates derived from the manually classified reference pixels. All tested algorithms were shown to be sensitive to overexposure. Three of the algorithms showed an accuracy which was high enough to be recommended for the processing of histogram-exposed hemispherical photographs: “Minimum” ( 98.8%; 0.952), “Edge Detection” ( 98.1%; 0.950), and “Minimum Histogram” ( 98.1%; 0.947). The Minimum algorithm overestimated gap fraction least of all (11%). The overestimation by the algorithms Edge Detection (63%) and Minimum Histogram (67%) were considerably larger. For the remaining four evaluated algorithms (IsoData, Maximum Entropy, MinError, and Otsu) an incompatibility with photographs containing overexposed pixels was detected. When applied to histogram-exposed photographs, these algorithms overestimated the gap fraction by at least 180%.     

Complete intra-annual cycle of leaf area index in a Platanus orientalis L. stand

Leaf area index (LAI) analysis of deciduous forest trees is usually restricted to seasonal monitoring involving the assessment of distinct leaf phenological stages within definite time intervals of the year. However, continuous LAI monitoring that includes entire leaf periods is necessary to define the ecophysiological characteristics of deciduous trees. Therefore, this study investigated the intra-annual cycle of the LAI for a Platanus orientalis L. stand in the Bart?n watershed of Turkey. A complete cycle involves three periods: foliation, stable, and defoliation. The foliation period comprises budburst, leaf emergence and flushing sessions, whereas the defoliation period consists of leaf senescence and leaf fall sessions. The stable period is in between these two periods when LAI values are at a climax around maximum. Eight points were determined in the field for the analysis of LAI by a hemispherical photography technique. Over a relatively frequent schedule, photographs were taken almost weekly during the foliation period. Both weekly and approximate monthly photographs were applied during the stable period. Finally, near-monthly photographs were taken for the defoliation period. The foliation period lasted for about 1.5 months from mid-April to May with the mean LAI reaching from 0.16 up to 2.38. Mean LAI was between 2.38 and 2.47 for a stable period over 2 months (June and July). For the defoliation period, mean LAI dropped from 2.42 down to 0.35 over 5 months from August to December. The total foliated period was more than 8 months, which is relatively long for a temperate forest. In addition, correlations between mean LAI and maximum, mean and minimum temperatures were highly significant (P < 0.01) with coefficients (r) of 0.79, 0.90 and 0.93, respectively. By describing the intra-annual LAI pattern, this study fills a gap in the literature on the phenology of Platanus orientalis L.     

Using terrestrial laser scanner for estimating leaf areas of individual trees in a conifer forest

A method that applies the terrestrial laser scanning to estimate leaf areas of individual trees in a mature conifer forest is presented. It is based on the theory of conventional optical LAI determinations, but refined for the inclusion of 3D depth information from the laser scanner. For each objective tree, we first used a single scan to measure local gap fractions beyond determined crown depths and combined this scan with other scans to delineate the geometrical dimensions of the crown. Then, through integrating the information from both aspects, the local leaf area density and the corresponding volume were derived. Finally, the total leaf area was obtained as the scalar product of these two variables. As most procedures were implemented on segmented 2D range images, the method possesses high efficiency. Additionally, through using gap fraction beyond determined crown depths, it solved the zero gap fraction problem encountered in segmented hemispherical photograph analysis. The method was tested on 11 trees in a 39 years old Norway spruce (Picea abies [L.] Karst.) stand located in southern Bavaria, Germany. Through correlation of the results with the estimates obtained with allometric equations, the accuracy was validated. The influence of the crown depth, for measuring gap fraction, and the segment size on estimation were also analyzed.     

基于PROSAIL辐射传输模型的毛竹林叶面积指数遥感反演

采用PROSAIL辐射传输模型建立毛竹林叶面积指数(LAI) 冠层反射率查找表,并结合Landsat TM卫星遥感数据,实现了毛竹林LAI的定量反演.结果表明： PROSAIL模型各输入参数的敏感性由高到低依次为LAI>叶绿素含量(C_ab)>叶片结构参数(N)>平均叶倾角(ALA)>等效水厚度(C_w)>干物质含量(C_m),并以LAI、C_ab两个主要敏感因子用于构建毛竹林LAI 冠层反射率查找表;基于PROSAIL模型的毛竹林LAI遥感反演结果与实测LAI具有很好的一致性,二者相关系数为0.90,均方根误差和相关的均方根误差也较小,分别为0.58和13.0%,但也存在反演LAI平均值高于实际值的问题.     

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.

     

Optimal nitrogen allocation controls tree responses to elevated CO2

Despite the abundance of experimental data, understanding of forest responses to elevated CO2 is limited. Here I show that a key to previously unexplained production and leaf area responses lies in the interplay between whole-plant nitrogen (N) allocation and leaf photosynthesis. A simple tree growth model, controlled by net growth maximization through optimization of leaf area index (LAI) and plant N, is used to analyse CO2 responses in both young, expanding and closed, steady-state canopies. The responses are sensitive to only two independent parameters, the photosynthetic capacity per leaf N (a) and the fine-root N:leaf N ratio. The model explains observed CO2 responses of photosynthesis, production and LAI in four forest free air CO2 enrichment (FACE) experiments. Insensitivity of LAI except at low LAI, increase in light-use efficiency, and photosynthetic down-regulation (as a result of reduced leaf N per area) at elevated CO2 are all explained through the combined effects on a and leaf quantum efficiency. The model bridges the gap between the understanding of leaf-level and plant-level responses and provides a transparent framework for interpreting and linking structural (LAI) and functional (net primary production (NPP):gross primary production (GPP) ratio, light-use efficiency, photosynthetic down-regulation) responses to elevated CO2.     

祁连山区青海云杉林冠层叶面积指数的反演方法

叶面积指数(Leaf area index, LAI)是陆地生态系统的一个十分重要的结构参数。随着空间精细化模型的发展和基于过程的分布式模拟技术的应用, 对LAI的区域估算显得越来越重要, 但目前尚缺乏有效的估算手段。该项研究以青海云杉(Picea crassifolia)林为研究对象, 利用LAI-2000冠层分析仪、鱼眼镜头法和经验公式法对林冠层LAI进行了测定, 观测值分别为1.03~3.70、0.48~2.26和2.27~8.20, 显然, 仪器测定值偏低。针对针叶的集聚效应导致仪器测定值偏低的现象, 利用跟踪辐射与冠层结构测量仪(TRAC)测定的青海云杉林聚集系数计算调整系数, 对鱼眼镜头法获取的LAI值进行订正。根据高分辨率的遥感数据反演青海云杉林的植被指数与LAI的关系, 最后获得了较合理的该地区林冠层LAI的空间分布图。     

暖温带地区3种森林群落叶面积指数和林冠开阔度的季节动态

以辽东栎(Quercus liaotungensis)为主的落叶阔叶林、华北落叶松(Larix principis-rupprechtii)林和油松(Pinus tabulaeformis)林是暖温带林区具有代表性的森林群落类型。该研究应用国内外流行的半球图方法,通过对这3种森林群落叶面积指数和林冠开阔度的测定和综合比较,分析了叶面积指数和林冠开阔度的季节动态,揭示了暖温带地区不同类型森林群落叶面积指数和林冠开阔度的特征。研究结果表明,落叶阔叶林(优势种为辽东栎、棘皮桦(Betula dahurica)和五角枫(Acer mono))和华北落叶松林两种落叶森林群落的叶面积指数值均随生长季的到来而呈现增长的趋势,最大值出现在8月;林冠开阔度值随着生长季的到来而下降,最大值出现在11月。落叶阔叶林的叶面积指数和林冠开阔度的季节动态较之华北落叶松林明显。油松是常绿树种,其群落叶面积指数和林冠开阔度的变化程度均不明显,但林冠开阔度的变化趋势也是与叶面积指数的变化趋势相反。通过计算得出叶面积指数和林冠开阔度相关显著,并且呈现指数回归的关系。此研究结果为以遥感途径获取暖温带地区叶面积指数提供了地面校正依据,为研究该地区植被林冠的异质性及其造成的影响,以及进一步对该地区林分、景观和区域尺度上碳、水分和通量等方面的模拟提供了基础数据。     

Comparing PAR transmission models for forest understorey vegetation

Question: Do Beer's Law models, multi‐layer scattering models, and a semi‐empirical model for predicting PAR transmission through understorey vegetation give comparable results? Do different driving variables (LAI, PLAI and percentage cover) give different results? How do the models vary when fit with species‐specific, species‐average and the ‘default’ parameters recommended in the literature? Location: Upland boreal forests of western North America. Methods: In calibration and validation plots, PAR transmission was measured, total cover visually estimated, and leaf dispersion, PLAI and cover estimated for each species using a point‐frame. Leaf inclination was measured by clinometer. PAR transmission was modelled using empirically‐fit Beer's Law models, a semi‐empirical model based on hemispherical gap fraction and first‐order scattering, and a multi‐layer model allowing multiple scattering. All models were modified to use leaf area index (LAI), vertically projected leaf area index (PLAI), or percentage cover data. Results: The empirical Beer's Law models had the least bias and best precision in predicting PAR transmission. The semi‐empirical model also had little bias and good precision, since the scattering coefficient compensated for problems in the estimation of gap fraction. The multi‐layer model consistently underestimated transmission. There was little benefit in accounting for species separately. LAI and PLAI‐based models were the most precise, but percentage cover models also provided reasonable predictions of PAR transmission. Conclusions: PAR transmission through forest understories can be simply modelled with Beer's Law using one empirical coefficient representing the average understorey species. More complex scattering models are less effective, likely because they fail to account for the complexity of the dispersion of this vegetation layer and its effect on radiation scattering.     

Calibration of LAI-2000 canopy analyser with leaf area index in a young eucalypt stand

Plant area index (PAI) measured with a LI-COR LAI-2000 plant canopy analyser (PCA) was calibrated with leaf area index (LAI) in a young stand of Eucalyptus grandis in the KwaZulu-Natal Midlands, South Africa. Destructive sampling and allometric equations were used to estimate LAI at 2 and 3 years after planting. Significant correlations (P<0.001) were found between LAI and PAI for each age with different equations being generated for the two ages (LAI=1.0594(PAI)−0.892 at 2 years of age, and LAI=1.0393(PAI) at 3 years of age). The equations differed from those reported in other eucalypt studies, as the PCA in this study over-predicted LAI at 2 years, and slightly under-predicted at 3 years, of age. It is argued that the stage of growth influenced this calibration, as the canopy and foliar structure may have been different in the young stands, affecting the basic assumptions for the PCA. A broad conversion from PCA derived PAI to LAI may not necessarily be valid for young, short rotation eucalypt plantations.     

Estimation of leaf area index and gap fraction in two broad-leaved forests by using small-footprint airborne LiDAR

In this study, we evaluated methods for reliably estimating leaf area index (LAI) and gap fraction in two different types of broad-leaved forests by the use of airborne light detection and ranging (LiDAR) data. We evaluated 13 estimation variables related to laser height, laser penetration rate, and laser point attributes that were derived from LiDAR analyses. The relationships between LiDAR-derived estimates and field-based measurements taken from the forests were evaluated with simple linear regressions. The data from the two forests were analyzed separately and as an integrated dataset. Among the laser height variables, the coefficient of variation (CV) of all laser point heights had the highest level of accuracy for estimating both LAI and gap fraction. However, we recommend that more evaluations be conducted prior to the use of CV in forests with complex structures. The simplest laser penetration variable, which represents the ratio of the number of ground points to the total number of all points (P _ALL), also had a high level of accuracy for estimating LAI and gap fraction at the study sites regardless of whether the data were analyzed separately or as an integrated data set. Furthermore, P _ALL values showed near 1:1 relationships with the field-based gap fraction values. Hence, the use of P _ALL may be the most practical for estimating LAI and gap fraction in broad-leaved forests, even when the canopies are heavily closed.     

林地叶面积指数遥感估算方法适用分析

叶面积指数是与森林冠层能量和CO2交换密切相关的一个重要植被结构参数,为了探讨估算林地叶面积指数LAI的遥感适用方法和提高精度的途径,利用TRAC仪器测定北京城区森林样地的LAI,从Landsat TM遥感图像计算NDVI、SR、RSR、SAVI植被指数,分别建立估算LAI的单植被指数统计模型、多植被指数组合的改进BP神经网络,获取最有效描述LAI与植被指数非线性关系的方法并应用到TM图像估算北京城区LAI。结果表明,单植被指数非线性统计模型估算LAI的精度高于线性统计模型;多植被指数组合神经网络中,以NDVI、RSR、SAVI组合估算LAI的精度最高,估算值与观测值线性回归方程的R2最高,为0.827,而RMSE最低,为0.189,神经网络解决了多植被指数组合统计模型非线性回归方程的系数较多、较难确定的问题,可较为有效的应用于遥感图像林地LAI的估算。     

杨树林全生长期LAI遥感估算模型适用性

基于时间序列的植被叶面积指数(LAI)估算方法一直是遥感领域研究的热点,对植被全生长期LAI进行估算以跟踪其生长情况具有重要的实用意义。以此为出发点,以滁州地区杨树林为研究对象,获取多时相环境卫星CCD(简称HJ-CDD)遥感影像,并利用LAI-2000同步测量杨树林叶面积指数(LAI)。使用归一化植被指数(NDVI)分别建立展叶期、花果期、叶面积稳定期和落叶始期的LAI估算模型,通过对比分析得到了全生长期LAI估算模型,并利用实测LAI对估算LAI进行了验证。最后进一步对该模型的适用性进行了验证,结果表明,此模型对于各个时期LAI的估算具有一定的适用性和有效性,可用于全生育期的遥感LAI生成,从而为LAI的动态变化监测提供了一种有效的研究思路和方法途径。     

Measuring leaf area index in rubber plantations − a challenge

In order to estimate water use, water requirements and carbon sequestration of tropical plantation systems such as rubber it is adamant to have accurate information on leaf area development of the plantation as the main determinant of evapotranspiration. Literature commonly suggests a number of different methods on how to obtain leaf area index (LAI) information from tree plantation systems. Methods include destructive measurements of leaf area at peak LAI, indirect methods such as gap fraction methods (i.e. Hemiview and LAI 2000) and radiation interception methods (i.e. SunScan) or litter fall traps. Published values for peak LAI in rubber plantation differ widely and show no clear trend to be explained by management practices or the influence of local climate patterns. This study compares four methods for determining LAI of rubber plantations of different ages in Xishuangbanna, Yunnan, PR China. We have tested indirect measurement techniques such as light absorption and gap fraction measurements and hemispherical image analysis against litter fall data in order to obtain insights into the reliability of these measuring techniques for the use in tropical tree plantation systems. In addition, we have included data from destructive harvesting as a comparison. The results presented here clearly showed that there was no consistent agreement between the different measurements. Site, time of the day and incoming radiation all had a significant effect on the results depending on the devices used. This leaves us with the conclusion that the integration of published data on LAI in rubber into broad ranging assessments is very difficult to accomplish as the accuracy of the measurements seems to be very sensitive to a number of factors. This diminishes the usefulness of literature data in estimating evapotranspiration from rubber plantations and the induced environmental effects, both on local as well as regional levels.     

Estimating leaf area index using unmanned aerial vehicle data: shallow vs. deep machine learning algorithms

Measuring leaf area index (LAI) is essential for evaluating crop growth and estimating yield, thereby facilitating high-throughput phenotyping of maize (Zea mays). LAI estimation models use multi-source data from unmanned aerial vehicles (UAVs), but using multimodal data to estimate maize LAI, and the effect of tassels and soil background, remain understudied. Our research aims to (1) determine how multimodal data contribute to LAI and propose a framework for estimating LAI based on remote-sensing data, (2) evaluate the robustness and adaptability of an LAI estimation model that uses multimodal data fusion and deep neural networks (DNNs) in single- and whole growth stages, and (3) explore how soil background and maize tasseling affect LAI estimation. To construct multimodal datasets, our UAV collected red–green–blue, multispectral, and thermal infrared images. We then developed partial least square regression (PLSR), support vector regression, and random forest regression models to estimate LAI. We also developed a deep learning model with three hidden layers. This multimodal data structure accurately estimated maize LAI. The DNN model provided the best estimate (coefficient of determination [R²] = 0.89, relative root mean square error [rRMSE] = 12.92%) for a single growth period, and the PLSR model provided the best estimate (R² = 0.70, rRMSE = 12.78%) for a whole growth period. Tassels reduced the accuracy of LAI estimation, but the soil background provided additional image feature information, improving accuracy. These results indicate that multimodal data fusion using low-cost UAVs and DNNs can accurately and reliably estimate LAI for crops, which is valuable for high-throughput phenotyping and high-spatial precision farmland management.

Multimodal data fusion (red–green–blue, multispectral, and thermal infrared) using low-cost unmanned aerial vehicles in a deep neural network and machine learning framework estimates maize leaf area index     

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.