Amazon Forest Structure from IKONOS Satellite Data and the Automated Characterization of Forest Canopy Properties

We developed an automated tree crown analysis algorithm using 1-m panchromatic IKONOS satellite images to examine forest canopy structure in the Brazilian Amazon. The algorithm was calibrated on the landscape level with tree geometry and forest stand data at the Fazenda Cauaxi (3.75° S, 48.37° W) in the eastern Amazon, and then compared with forest stand data at Tapajos National Forest (3.08° S, 54.94° W) in the central Amazon. The average remotely sensed crown width (mean ± SE) was 12.7 ± 0.1 m (range: 2.0–34.0 m) and frequency of trees was 76.6 trees/ha at Cauaxi. At Tapajos, remotely sensed crown width was 13.1 ± 0.1 m (range: 2.0–38.0 m) and frequency of trees was 76.4 trees/ha. At both Cauaxi and Tapajos, the remotely sensed average crown widths were within 3 percent of the crown widths derived from field measurements, although crown distributions showed significant differences between field-measured and automated methods. We used the remote sensing algorithm to estimate crown dimensions and forest structural properties in 51 forest stands (1 km²) throughout the Brazilian Amazon. The estimated crown widths, tree diameters (dbh), and stem frequencies differed widely among sites, while estimated biomass was similar among most sites. Sources of observed errors included an inability to detect understory crowns and to separate adjacent, intermingled crowns. Nonetheless, our technique can serve to provide information about structural characteristics of large areas of unsurveyed forest throughout Amazonia.     

Allometric equations for integrating remote sensing imagery into forest monitoring programmes

Remote sensing is revolutionizing the way we study forests, and recent technological advances mean we are now able – for the first time – to identify and measure the crown dimensions of individual trees from airborne imagery. Yet to make full use of these data for quantifying forest carbon stocks and dynamics, a new generation of allometric tools which have tree height and crown size at their centre are needed. Here, we compile a global database of 108753 trees for which stem diameter, height and crown diameter have all been measured, including 2395 trees harvested to measure aboveground biomass. Using this database, we develop general allometric models for estimating both the diameter and aboveground biomass of trees from attributes which can be remotely sensed – specifically height and crown diameter. We show that tree height and crown diameter jointly quantify the aboveground biomass of individual trees and find that a single equation predicts stem diameter from these two variables across the world's forests. These new allometric models provide an intuitive way of integrating remote sensing imagery into large‐scale forest monitoring programmes and will be of key importance for parameterizing the next generation of dynamic vegetation models.     

Age‐related Crown Thinning in Tropical Forest Trees

Gap dynamics theory proposes that treefall gaps provide high light levels needed for regeneration in the understory, and by increasing heterogeneity in the light environment allow light‐demanding tree species to persist in the community. Recent studies have demonstrated age‐related declines in leaf area index of individual temperate trees, highlighting a mechanism for gradual changes in the forest canopy that may also be an important, but less obvious, driver of forest dynamics. We assessed the prevalence of age‐related crown thinning among 12 tropical canopy tree species sampled in lowland forests in Panama and Puerto Rico (total N = 881). Canopy gap fraction of individual canopy tree crowns was positively related to stem diameter at 1.3 m (diameter at breast height) in a pooled analysis, with 10 of 12 species showing a positive trend. Considered individually, a positive correlation between stem diameter and canopy gap fraction was statistically significant in 4 of 12 species, all of which were large‐statured canopy to emergent species: Beilschmiedia pendula, Ceiba pentandra, Jacaranda copaia, and Prioria copaifera. Pooled analyses also showed a negative relationship between liana abundance and canopy gap fraction, suggesting that lianas could be partially obscuring age‐related crown thinning. We conclude that age‐related crown thinning occurs in tropical forests, and could thus influence patterns of tree regeneration and tropical forest community dynamics.     

Vegetation Structure,Composition, and Species Richness Across a 56‐year Chronosequence of Dry Tropical Forest on Providencia Island,Colombia1

We compared vegetation structure and species richness across a 56‐yr chronosequence of six replicated age classes of dry tropical forest on the island of Providencia, Colombia, in the Southwest Caribbean. Stand age classes were determined using sequential, orthorectified panchromatic aerial photos acquired between 1944 and 1996 and Landsat 7 ETM + satellite imagery from 2000. Along the chronosequence we established 59 plots of 2 × 50 m (0.01 ha) to document changes in species richness, basal area, tree height, stem density, and sprouting ability. All woody trees and shrubs >2.5 cm diameter at breast height (DBH) were censused and measured. Although woody species density reached a peak in stands from 32 to 56 yr old, rarefaction analysis showed that species richness increased linearly with stand age and was highest in stands 56 yr old or greater. Nonparametric, abundance‐based estimators of species richness also showed positive and linear associations with age. Basal area and mean tree height were positively associated with age since abandonment, while sprouting ability showed a negative relationship. Our results indicate rapid recovery of woody species richness and structural characteristics along this tropical dry forest chronosequence.     

Developmental Patterns of Tree Dimensions in a Neotropical Deciduous Forest1

For 26 tree species in very dry tropical forest in Mexico, the developmental trends of relationships among trunk diameter, tree height, and crown diameter were inferred from a one‐time measurement of dispersed individuals across the size range from saplings to large, mature trees. On hillside sites in this high diversity forest, maximum dimensions were usually <10‐m height, 4‐m crown diameter, and 0.3‐m trunk diameter. The relationship of height to trunk diameter was characterized by an asymptotic, three‐parameter model. Crown diameter was a linear function of trunk diameter. The parameter values for both models varied widely among the species. In general, the dispersion among species of the height–crown diameter relationship increased linearly with trunk diameter (up to 0.2 m). Arborescent cacti were distant from other species at all sizes, although they were well modeled using the same equations. Empirical and theoretical features and limitations of the present and previous models, including mechanical buckling and water‐stress theories, are considered.     

Quantifying mortality of tropical rain forest trees using high-spatial-resolution satellite data

Assessment of forest responses to climate change is severely hampered by the limited information on tree death on short temporal and broad spatial scales, particularly in tropical forests. We used 1‐m resolution panchromatic IKONOS and 0.7‐m resolution QuickBird satellite data, acquired in 2000 and 2002, respectively, to evaluate tree death rates at the La Selva Biological Station in old‐growth Tropical Wet Forest in Costa Rica, Central America. Using a calibration factor derived from ground inspection of tree deaths predicted from the images, we calculated a landscape‐scale annual exponential death rate of 2.8%. This corresponds closely to data for all canopy‐level trees in 18 forest inventory plots, each of 0.5 ha, for a mostly‐overlapping 2‐year period (2.8% per year). This study shows that high‐spatial‐resolution satellite data can now be used to measure old‐growth tropical rain forest tree death rates, suggesting many new avenues for tropical forest ecology and global change research.     

Ontogenetic changes in size, allometry, and mechanical design of tropical rain forest trees

Size, allometry, and mechanical design were measured for trees of three canopy species in a tropical rain forest in French Guiana. Mechanical design was expressed as the safety factor, using the elastic-stability model, and the wind resistance factor, using the constant-stress model. Changes with ontogeny were described as regressions using stem diameter as the independent variable, and they were compared between species. Height, crown size, and the wind resistance factor increased with ontogeny. The safety factor decreased to a minimum and then increased continuously in thicker trees. The crown width/height ratio did not change with ontogeny. Interspecific differences in allometry and mechanical design were related to the adult stature of the species, and not to shade tolerance. The short stature species (Vouacapoua americana) was less slender (height:DBH [stem diameter at 1.3 m] ratio) and had a higher crown width/height ratio than the tall stature species (Goupia glabra and Dicorynia guianensis). Vouacapoua had a higher safety factor, but a similar wind resistance factor. The safety factors of our study species were lower than those of two temperate tree species because of a higher slenderness. Differences in safety factors between tropical and temperate trees may result from unrealistic assumptions of the elastic-stability model, and may also be related to lower light levels and-or wind rates in the tropics.     

Assessing the above‐ground biomass of a complex tropical rainforest using a canopy crane

Abstract Current estimates of the total biomass in tropical rainforests vary considerably; this is due in large part to the different approaches that are used to calculate biomass. In this study we have used a canopy crane to measure the tree architectures in a 1 ha plot of complex mesophyll vine forest at Cape Tribulation, Australia. Methods were developed to measure and calculate the crown and stem biomass of six major species of tree and palm (Alstonia scholaris (Apocynaceae), Cleistanthus myrianthus (Euphorbiaceae), Endiandra microneura (Lauraceae), Myristica insipida (Myristicaceae), Acmena graveolens (Myrtaceae), Normanbya normanbyi (Arecaceae)) using the unique access provided by the crane. This has allowed the first non‐destructive biomass estimate to be carried out for a forest of this type. Allometric equations which relate tree biomass to the measured variable ‘diameter at breast height’ were developed for the six species, and a general equation was also developed for trees on the plot. The general equation was similar in form to equations developed for tropical rainforests in Brazil and New Guinea. The species equations were applied at the level of families, the generalized equation was applied to the remaining species which allowed the biomass of a total of 680 trees to be calculated. This has provided a current estimate of 270 t ha⁻¹ above‐ground biomass at the Australian Canopy Crane site; a value comparable to lowland rainforests in Panama and French Guiana. Using the same tree database seven alternative allometric equations (literature equations for tropical rainforests) were used to calculate the site biomass, the range was large (252–446 t ha⁻¹) with only three equations providing estimates within 34 t ha⁻¹ (12.5%) of the site value. Our use of multiple species‐specific allometric equations has provided a site estimate only slightly larger (1%) than that obtained using allometric equations developed specifically for tropical wet rainforests. We have demonstrated that it is possible to non‐destructively measure the biomass in a complex forest using an on‐site canopy crane. In conjunction the development of crown maps and a detailed tree architecture database allows changes in forest structure to be followed quantitatively.     

Depth of soil water uptake by tropical rainforest trees during dry periods: does tree dimension matter?

Though the root biomass of tropical rainforest trees is concentrated in the upper soil layers, soil water uptake by deep roots has been shown to contribute to tree transpiration. A precise evaluation of the relationship between tree dimensions and depth of water uptake would be useful in tree-based modelling approaches designed to anticipate the response of tropical rainforest ecosystems to future changes in environmental conditions. We used an innovative dual-isotope labelling approach (deuterium in surface soil and oxygen at 120-cm depth) coupled with a modelling approach to investigate the role of tree dimensions in soil water uptake in a tropical rainforest exposed to seasonal drought. We studied 65 trees of varying diameter and height and with a wide range of predawn leaf water potential (Ψ_pd) values. We confirmed that about half of the studied trees relied on soil water below 100-cm depth during dry periods. Ψ_pd was negatively correlated with depth of water extraction and can be taken as a rough proxy of this depth. Some trees showed considerable plasticity in their depth of water uptake, exhibiting an efficient adaptive strategy for water and nutrient resource acquisition. We did not find a strong relationship between tree dimensions and depth of water uptake. While tall trees preferentially extract water from layers below 100-cm depth, shorter trees show broad variations in mean depth of water uptake. This precludes the use of tree dimensions to parameterize functional models.     

Decaying trees improve nesting opportunities for cavity‐nesting birds in temperate and boreal forests: A meta‐analysis and implications for retention forestry

Many studies have dealt with the habitat requirements of cavity‐nesting birds, but there is no meta‐analysis on the subject and individual study results remain vague or contradictory. We conducted a meta‐analysis to increase the available evidence for nest‐site selection of cavity‐nesting birds. Literature was searched in Web of Science and Google Scholar and included studies that provide data on the habitat requirements of cavity‐nesting birds in temperate and boreal forests of varying naturalness. To compare nest and non‐nest‐tree characteristics, the following data were collected from the literature: diameter at breast height (DBH) and its standard deviation (SD), sample size of trees with and without active nest, amount of nest and available trees described as dead or with a broken crown, and amount of nest and available trees that were lacking these characteristics. Further collected data included bird species nesting in the cavities and nest‐building type (nonexcavator/excavator), forest type (coniferous/deciduous/mixed), biome (temperate/boreal), and naturalness (managed/natural). From these data, three effect sizes were calculated that describe potential nest trees in terms of DBH, vital status (dead/alive), and crown status (broken/intact). These tree characteristics can be easily recognized by foresters. The results show that on average large‐diameter trees, dead trees, and trees with broken crowns were selected for nesting. The magnitude of this effect varied depending primarily on bird species and the explanatory variables forest type and naturalness. Biome had lowest influence (indicated by ΔAIC). We conclude that diameter at breast height, vitality, and crown status can be used as tree characteristics for the selection of trees that should be retained in selectively harvested forests.     

Sleeping site preferences in tufted capuchin monkeys (Cebus apella nigritus)

The characteristics and availability of the sleeping sites used by a group of 27 tufted capuchin monkeys (Cebus apella nigritus) were studied during 17 months at the Iguazu National Park, Argentina. We tested different hypotheses regarding possible ultimate causes of sleeping-site selection. Most sleeping sites were located in areas of tall, mature forest. Of the 34 sleeping sites the monkeys used during 203 nights, five were more frequently used than the others (more than 20 times each, constituting 67% of the nights). Four species of tree (Peltophorum dubium, Parapiptadenia rigida, Copaifera langsdorfii and Cordia trichotoma) were the most frequently used. They constituted 82% of all the trees used, though they represent only 12% of the trees within the monkeys' home range which had a diameter at breast height (DBH) > 48.16 cm (1 SD below the mean DBH of sleeping trees). The sleeping trees share a set of characteristics not found in other trees: they are tall emergent (mean height +/- SD = 31.1+/-5.2 m) with large DBH (78.5+/-30.3 cm), they have large crown diameter (14+/-5.5 m), and they have many horizontal branches and forks. Adult females usually slept with their kin and infants, while peripheral adult males sometimes slept alone in nearby trees. We reject parasite avoidance as an adaptive explanation for the pattern of sleeping site use. Our results and those from other studies suggest that predation avoidance is a predominant factor driving sleeping site preferences. The patterns of aggregation at night and the preference for trees with low probability of shedding branches suggest that social preferences and safety from falling during windy nights may also affect sleeping tree selection. The importance of other factors, such as seeking comfort and maintaining group cohesion, was not supported by our results. Other capuchin populations show different sleeping habits which can be explained by differences in forest structure and by demographic differences.     

Crown allometries are less responsive than stem allometry to tree size and habitat variations in an Indian monsoon forest

While theoretical allometric models postulate universal scaling exponents, empirical relationships between tree dimensions show marked variability that reflects changes in the biomass allocation pattern. As growth of the various tree compartments may be controlled by different functions, it is hypothesized that they may respond differently to factors of variation, resulting in variable tree morphologies and potentially in trade-offs between allometric relationships. We explore the variability of tree stem and crown allometries using a dataset of 1,729 trees located in an undisturbed wet evergreen forest of the Western Ghats, India. We specifically test whether species adult stature, terrain slope, tree size and crown light exposure affect the relationships between stem diameter and stem height (stem allometry), and between stem diameter and crown width, crown area and crown volume (crown allometries). Results show that both stem and crown allometries are subject to variations in relation to both endogenous (tree size, species adult stature) and exogenous (terrain slope, crown light exposure) factors. Stem allometry appears to be more affected by these factors than are crown allometries, including the stem diameter–crown volume relationship, which proved to be particularly stable. Our results support the idea that height is a prevailing adjustment factor for a tree facing variable growth (notably light) conditions, while stem diameter–crown volume allometry responds more to internal metabolic constraints. We ultimately discuss the various sources of variability in the stem and crown allometries of tropical trees that likely play an important role in forest community dynamics.     

Validating Community-Led Forest Biomass Assessments

The lack of capacity to monitor forest carbon stocks in developing countries is undermining global efforts to reduce carbon emissions. Involving local people in monitoring forest carbon stocks could potentially address this capacity gap. This study conducts a complete expert remeasurement of community-led biomass inventories in remote tropical forests of Papua New Guinea. By fully remeasuring and isolating the effects of 4,481 field measurements, we demonstrate that programmes employing local people (non-experts) can produce forest monitoring data as reliable as those produced by scientists (experts). Overall, non-experts reported lower biomass estimates by an average of 9.1%, equivalent to 55.2 fewer tonnes of biomass ha^-1, which could have important financial implications for communities. However, there were no significant differences between forest biomass estimates of expert and non-expert, nor were there significant differences in some of the components used to calculate these estimates, such as tree diameter at breast height (DBH), tree counts and plot surface area, but were significant differences between tree heights. At the landscape level, the greatest biomass discrepancies resulted from height measurements (41%) and, unexpectedly, a few large missing trees contributing to a third of the overall discrepancies. We show that 85% of the biomass discrepancies at the tree level were caused by measurement taken on large trees (DBH ≥50cm), even though they consisted of only 14% of the stems. We demonstrate that programmes that engage local people can provide high-quality forest carbon data that could help overcome barriers to reducing forest carbon emissions in developing countries. Nonetheless, community-based monitoring programmes should prioritise reducing errors in the field that lead to the most important discrepancies, notably; overcoming challenges to accurately measure large trees.     

樟子松人工林一级枝条基径和枝长模型的研究

以东北林业大学帽儿山实验林场樟子松人工林为研究对象,采用枝解析的方法,于2002年、2003年测定了53株林木（年龄17~38 a,直径8.61~21.5 cm,树高7.48~18.24 m）的树冠变量,建立了基于总着枝深度（DINC）和树冠内一级枝条基径（BD）、枝长（BL）的预估模型。对于大小相同树木的一级枝条, 这些树冠变量随着DINC的增加而增大,而林木的胸径（DBH）、树高（HT）变量又很好地反映了不同大小树木的基径和枝长的变化。采用独立检验样本对构建的树冠内一级枝条基径和枝长模型进行了拟合统计量和精度检验,结果表明：模型预测效果良好,精度均达到95%以上。构建的一级枝条基径和枝长模型为进一步合理地描述樟子松人工林树冠的形状及其变化以及三维可视化经营提供依据。     

Estimating Stem Volume Using QuickBird Imagery and Allometric Relationships for Open Populus xiaohei Plantations

There has been a great deal of interest in studying the crown of trees using remote sensing data. In this study, crown width was extracted from high-resolution QuickBird images for open Populus xiaohei plantations. Regression models for predicting the individual stem volumes of Populus xiaohei were established using extracted crown width, as well as estimated tree parameters （i.e. diameter at breast height [DBH] and tree height） as predictors. Our results indicated that crown width could be accurately extracted from QuickBird images using a multi-scale segmentation approach with a mean relative error of 5.74%, especially for wide-spacing stands. Using either extracted crown width alone or with estimated DBH and tree height can successfully estimate individual stem volume of Populus xiaohei with the R2 value ranging from 0.87 to 0.92 depending on different model forms. In particular, the two second-order polynomial models （model 2 and model 6）, based on QuickBird image-derived crown widths and estimated DBH and tree heights, respectively, were the best at describing the relationship between stem volume and tree characteristics.     

Estimating Stem Volume Using QuickBird Imagery and Allometric Relationships for Open Populus xiaohei Plantations

There has been a great deal of interest in studying the crown of trees using remote sensing data.In this study,crownwidth was extracted from high-resolution QuickBird images for open Populus xiaohei plantations.Regression modelsfor predicting the individual stem volumes of Populus xiaohei were established using extracted crown width,as well asestimated tree parameters(i.e.diameter at breast height[DBH]and tree height)as predictors.Our results indicated thatcrown width could be accurately extracted from QuickBird images using a multi-scale segmentation approach with a meanrelative error of 5.74%,especially for wide-spacing stands.Using either extracted crown width alone or with estimatedDBH and tree height can successfully estimate individual stem volume of Populus xiaohei with the R~2 value ranging from0.87 to 0.92 depending on different model forms.In particular,the two second-order polynomial models(model2 andmodel 6),based on QuickBird image-derived crown widths and estimated DBH and tree heights,respectively,were the bestat describing the relationship between stem volume and tree characteristics.     

Viewing tropical forest succession as a three-dimensional dynamical system

As tropical forests are complex systems, they tend to be modelled either roughly via scaling relationships or in a detailed manner as high-dimensional systems with many variables. We propose an approach which lies between the two whereby succession in a tropical forest is viewed as a trajectory in the configuration space of a dynamical system with just three dependent variables, namely, the mean leaf-area index (LAI) and its standard deviation (SD) or coefficient of variation along a transect, and the mean diameter at breast height (DBH) of trees above the 90th percentile of the distribution of tree DBHs near the transect. Four stages in this forest succession are identified: (I) naturally afforesting grassland: the initial stage with scattered trees in grassland; (II) very young forest: mostly covered by trees with a few remaining gaps; (III) young smooth forest: almost complete cover by trees of mostly similar age resulting in a low SD; and (IV) old growth or mature forest: the attracting region in configuration space characterized by fluctuating SD from tree deaths and regrowth. High-resolution LAI measurements and other field data from Khao Yai National Park, Thailand show how the system passes through these stages in configuration space, as do simple considerations and a crude cellular automaton model.     

Age and Long-term Growth of Trees in an Old-growth Tropical Rain Forest, Based on Analyses of Tree Rings and 14C1

In an old‐growth tropical wet forest at La Selva, Costa Rica, we combined radiocarbon (¹⁴C) dating and tree‐ring analysis to estimate the ages of large trees of canopy and emergent species spanning a broad range of wood densities and growth rates. We collected samples from the trunks of 29 fallen, dead individuals. We found that all eight sampled species formed visible growth rings, which varied considerably in distinctiveness. For five of the six species for which we combined wood anatomical studies with ¹⁴C‐dates (ring ages), the analyses demonstrated that growth rings were of annual formation. The oldest tree we found by direct ring counting was a Hymenolobium mesoamericanum Lima (Papilionaceae) specimen, with an age of ca. 530 years at the time of death. All other sampled individuals, including very large trees of slow‐growing species, had died at ages between 200 and 300 years. These results show that, even in an everwet tropical rain forest, tree growth of many species can be rhythmic, with an annual periodicity. This study thus raises the possibility of extending tree‐ring analyses throughout the tropical forest types lacking a strong dry season or annual flooding. Our findings and similar measurements from other tropical forests indicate that the maximum ages of tropical emergent trees are unlikely to be much greater than 600 years, and that these trees often die earlier from various natural causes.     

Estimating the heights and diameters at breast height of trees in an urban park and along a street using mobile LiDAR

Because trees can positively influence local environments in urban ecosystems, it is important to measure their morphological characteristics, such as height and diameter at breast height (DBH). However, measuring these data for each individual tree is a time-consuming process that requires a great deal of manpower. In this study, we investigated the feasibility of using mobile LiDAR to estimate tree height and DBH along urban streets and in urban parks. We compared measurements from a mobile LiDAR unit with field measurements of tree height and DBH in urban parks and streets. The height-above-ground and Pratt circle fit methods were applied to calculate tree height and DBH, respectively. The LiDAR-estimated tree heights were highly accurate albeit slightly underestimated, with a root mean square error of 0.359 m for the street trees and 0.462 m for the park trees. On the other hand, the estimated DBHs were moderately accurate and overestimated, with a root mean square error of 3.77 cm for the street trees and 8.95 cm for the park trees. Densely planted trees in the park and obstacles in urban areas result in “shadows” (areas with no data), reducing accuracy. Irregular trunk shapes and scanned data that did not include full data point coverage of every trunk were the reasons for the errors. Despite these errors, this study highlights the potential of tree measurements obtained with mobile terrestrial LiDAR platforms to be scaled up from point-based locations to neighborhood-scale and city-scale inventories.

     

Hazard ratings of pine forests to a pine wilt disease at two spatial scales (individual trees and stands) using self-organizing map and random forest

Pine wilt disease (PWD) caused by the pine wood nematode is the most serious global threat to pine forests. Hazard ratings of trees and forests to pest attacks provide important information to efficiently identify current or future hazardous conditions. However, in spite of the importance of hazard ratings for managing PWD, there are few studies on hazard ratings in this system. In this study, we evaluated the hazard ratings of pine trees and pine stands to PWD by considering environmental factors at the level of the stand and the individual tree. Our results showed that trees with larger diameter at breast height (DBH) showed a higher risk rate than those with smaller DBH, indicating that large trees have an increased probability of exposure to vector beetles because they are tall and have a large crown volume. We also found that reduced tree vigour could be related to susceptibility to PWD. In pine stands, geographical factors showed a high correlation with the occurrence of PWD. PWD occurrence was rare at high altitudes, but was more common on steep and south-facing slopes. These patterns were consistently observed in the results from 2 computational approaches: self-organizing map (SOM) and random forest models. The combination of SOM and random forest was effective to extract ecological information from the dataset. The SOM efficiently characterized relations among variables, and the random forest model was effective at predicting ecological variables, including the hazard rating of trees to disturbances.