Search

Large‐scale dieback of ash trees (Fraxinus spp.) caused by the fungus Hymenoscyphus fraxineus is posing an immense threat to forest health in Europe, requiring effective monitoring at large scales. In this study, a pipeline was created to find ash trees and classify dieback severity using high‐resolution hyperspectral imagery of individual tree crowns (ITCs). Hyperspectral data were collected in four forest sites near Cambridge, UK, where 422 ITCs were manually delineated and labelled using field‐measurements of species and dieback severity (for ash trees). Four algorithms, namely linear discriminant analysis (LDA), principal components analysis coupled with LDA (PCA‐LDA), partial least squares discriminant analysis (PLS‐DA) and random forest (RF), were used to build classification models for species and dieback severity classification. The effect of dark‐pixel filtering on classification accuracy was evaluated. The best performing models were then coupled with automatic ITC segmentation to map species and ash dieback distribution over 16.8 hectares of woodland. We calculated and partitioned the coefficient of variation (CV) of the reflected ash spectra to find variable wavebands associated with dieback. PLS‐DA and LDA were most accurate for classifying ITC species identifies (overall accuracy >90%), whereas RF was most accurate for classifying ash dieback severity (overall accuracy 77%). Dark pixel filtering further increased the accuracy of species classification (+6%), but not disease classification. The reflectances of narrow blue (415 nm), red‐edge (680 nm) and NIR (760 nm) bands had high CV across disease classes and should be included if multispectral imagery were to be used to monitor ash dieback. The study demonstrates the possibility of using remote sensing to forward epidemiological research by monitoring forest pathogens in landscape scales, which would allow temperate forest managers to control pathogen outbreaks, assess associated impacts and restore affected forests much more effectively.

Tropical forests are subject to diverse deforestation pressures but their conservation is essential to achieve global climate goals. Predicting the location of deforestation is challenging due to the complexity of the natural and human systems involved but accurate and timely forecasts could enable effective planning and on-the-ground enforcement practices to curb deforestation rates. New computer vision technologies based on deep learning can be applied to the increasing volume of Earth observation data to generate novel insights and make predictions with unprecedented accuracy.Here, we demonstrate the ability of deep convolutional neural networks to learn spatiotemporal patterns of deforestation from a limited set of freely available global data layers, including multispectral satellite imagery, the Hansen maps of historic deforestation (2001-2020) and the ALOS JAXA digital surface model, to forecast future deforestation (2021). We designed four original deep learning model architectures, based on 2D Convolutional Neural Networks (2DCNN), 3D Convolutional Neural Networks (3DCNN), and Long Short-Term Memory (LSTM) Recurrent Neural Networks (RNN) to produce spatial maps that indicate the risk to each forested pixel (~30 m) in the landscape of becoming deforested within the next year. They were trained and tested on data from two ~80,000 km2 tropical forest regions in the Southern Peruvian Amazon.We found that the networks could predict the likely location of future deforestation to a high degree of accuracy. Our best performing model – a 3DCNN – had the highest pixel-wise accuracy (80-90%) when validated on 2020 deforestation based 2014-2019 training. Visual examination of the forecasts indicated that the 3DCNN network could automatically discern the drivers of forest loss from the input data. For example, pixels around new access routes (e.g. roads) were assigned high risk whereas this was not the case for recent, concentrated natural loss events (e.g. remote landslides).CNNs can harness limited time-series data to predict near-future deforestation patterns, an important step in using the growing volume of satellite remote sensing data to curb global deforestation. The modelling framework can be readily applied to any tropical forest location and used by governments and conservation organisations to prevent deforestation and plan protected areas.

Tropical forests are a major component of the global carbon cycle and home to two-thirds of terrestrial species. Upper-canopy trees store the majority of forest carbon and can be vulnerable to drought events and storms. Monitoring their growth and mortality is essential to understanding forest resilience to climate change, but in the context of forest carbon storage, large trees are underrepresented in traditional field surveys, so estimates are poorly constrained. Aerial photographs provide spectral and textural information to discriminate between tree crowns in diverse, complex tropical canopies, potentially opening the door to landscape monitoring of large trees. Here we describe a new deep convolutional neural network method,Detectree2, which builds on the Mask R-CNN computer vision framework to recognise the irregular edges of individual tree crowns from airborne RGB imagery. We trained and evaluated this model with 3,797 manually delineated tree crowns at three sites in Malaysian Borneo and one site in French Guiana. As an example application, we combined the delineations with repeat lidar surveys (taken between 3 and 6 years apart) of the four sites to estimate the growth and mortality of upper-canopy trees.Detectree2delineated 65,000 upper-canopy trees across 14 km2of aerial images. The skill of the automatic method in delineating unseen test trees was good (F1score = 0.64) and for the tallest category of trees was excellent (F1score = 0.74). As predicted from previous field studies, we found that growth rate declined with tree height and tall trees had higher mortality rates than intermediate-size trees. Our approach demonstrates that deep learning methods can automatically segment trees in widely accessible RGB imagery. This tool (provided as an open-source Python package) has many potential applications in forest ecology and conservation, from estimating carbon stocks to monitoring forest phenology and restoration.Python package available to install athttps://github.com/PatBall1/Detectree2

Peer reviewed: True, Funder: Bill & Melinda Gates Foundation, Burn-area products from remote sensing provide the backbone for research in fire ecology, management, and modelling. Landsat imagery could be used to create an accurate burn-area map time series at ecologically relevant spatial resolutions. However, the low temporal resolution of Landsat has limited its development in wet tropical and subtropical regions due to high cloud cover and rapid burn-area revegetation. Here, we describe a 34-year Landsat-based burn-area product for wet, subtropical Hong Kong. We overcame technical obstacles by adopting a new LTS fire burn-area detection pipeline that (1) Automatically uniformized Landsat scenes by weighted histogram matching; (2) Estimated pixel resemblance to burn areas based on a random forest model trained on the number of days between the fire event and the date of burn-area detection; (3) Iteratively merged features created by thresholding burn-area resemblance to generate burn-area polygons with detection dates; and (4) Estimated the burn severity of burn-area pixels using a time-series compatible approach. When validated with government fire records, we found that the LTS fire product carried a low area of omission (11%) compared with existing burn-area products, such as GABAM (49%), MCD64A1 (72%), and FireCCI51 (96%) while effectively controlling commission errors. Temporally, the LTS fire pipeline dated 76.9% of burn-area polygons within two months of the actual fire event. The product represents the first Landsat-based burn-area product in wet tropical and subtropical Asia that covers the entire time series. We believe that burn-area products generated from algorithms like LTS fire will effectively bridge the gap between remote sensing and field-based studies on wet tropical and subtropical fire ecology.

Field spectroscopy is a powerful tool for monitoring leaf functional traits in situ, but it remains unclear whether universal statistical models can be developed to predict traits from spectral information, or whether re-calibration is necessary as conditions vary. In particular, multiple leaf traits vary simultaneously across growing seasons, and it is an open question whether these temporal changes can be predicted successfully from hyperspectral data. To explore this question, monthly changes in 21 physiochemical leaf traits and plant spectra were measured for eight deciduous tree species from the UK. Partial least-squares regression (PLSR) was used to evaluate whether each trait could be predicted from a single PLSR model from reflectance spectra, or whether species- and month-level models were needed. Physiochemical traits and spectra varied greatly over the growing season, although there was less variation among mature leaves harvested between June and September. Importantly, leaf spectroscopy was able to predict seasonal variations of most leaf traits accurately, with accuracies of prediction generally higher for mature leaves. However, for several traits, the PLSR estimation models varied among species, and a single PLSR model could not be used to make accurate species-level predictions. Our findings demonstrate that leaf spectra can successfully predict multiple functional foliar traits through the growing season, establishing one of the fundamentals for monitoring and mapping plant functional diversity in temperate forests from air- and spaceborne imaging spectroscopy.

Hyperspectral images taken from aircraft or satellites contain information from hundreds of spectral bands, within which lie latent lower-dimensional structures that can be exploited for classifying vegetation and other materials. A disadvantage of working with hyperspectral images is that, due to an inherent trade-off between spectral and spatial resolution, they have a relatively coarse spatial scale, meaning that single pixels may correspond to spatial regions containing multiple materials. This article introduces the Diffusion and Volume maximization-based Image Clustering (D-VIC) algorithm for unsupervised material clustering to address this problem. By directly incorporating pixel purity into its labeling procedure, D-VIC gives greater weight to pixels that correspond to a spatial region containing just a single material. D-VIC is shown to outperform comparable state-of-the-art methods in extensive experiments on a range of hyperspectral images, including land-use maps and highly mixed forest health surveys (in the context of ash dieback disease), implying that it is well-equipped for unsupervised material clustering of spectrally-mixed hyperspectral datasets., Comment: 28 pages, 11 figures

The application of metabarcoding to environmental and invertebrate‐derived DNA (eDNA and iDNA) is a new and increasingly applied method for monitoring biodiversity across a diverse range of habitats. This approach is particularly promising for sampling in the biodiverse humid tropics, where rapid land‐use change for agriculture means there is a growing need to understand the conservation value of the remaining mosaic and degraded landscapes. Here we use iDNA from blood‐feeding leeches (Haemadipsa picta) to assess differences in mammalian diversity across a gradient of forest degradation in Sabah, Malaysian Borneo. We screened 557 individual leeches for mammal DNA by targeting fragments of the 16S rRNA gene and detected 14 mammalian genera. We recorded lower mammal diversity in the most heavily degraded forest compared to higher quality twice logged forest. Although the accumulation curves of diversity estimates were comparable across these habitat types, diversity was higher in twice logged forest, with more taxa of conservation concern. In addition, our analysis revealed differences between the community recorded in the heavily logged forest and that of the twice logged forest. By revealing differences in mammal diversity across a human‐modified tropical landscape, our study demonstrates the value of iDNA as a non‐invasive biomonitoring approach in conservation assessments.

Aim: Small, old-growth forest fragments generally have more small-seeded plants than large patches, due to the disappearance of large seed dispersing vertebrates. This pattern may differ for secondary forest fragments where differential migration ability rather than persistence of seed dispersers may be driving plant community assembly. In this study, we investigated the effect of habitat fragmentation on seed dispersers and plant community structure in regenerating forests. Location: The Thousand Island Lake, China. Taxon: Plants, birds and mammals. Methods: We compiled diversity and abundance data for birds and mammals on islands in the Thousand Island Lake, China. We also surveyed the secondary plant communities and measured seed dispersal traits. Results: Community-weighted mean seed size of woody plants decreased with island size. This pattern was related to compositional difference of the dispersers. We found that mammal diversity and abundance was only weakly or not related to island size; whereas bird diversity and abundance increased strongly with island size. Density of bird-dispersed plants was significantly positively related with island size. Since birds tend to disperse smaller seeds than mammals, the trend in seed size may have been a consequence of the shift in relative abundance of the two disperser guilds. Main Conclusions: Differential responses of seed dispersers to habitat fragmentation may lead to pervasive shifts in the plant community structure of regenerating forest fragments. Our study highlights the importance of keeping large continuous forests in order to retain mammals and their dispersal capabilities., National Natural Science Foundation of China. Grant Numbers: 31500382, 31570524, 31361123001, 31770462 and 31210103908 National Science Foundation. Grant Number: DEB‐1342754 and DEB‐134275

Reliable assessment of forest structural types (FSTs) aids sustainable forest management. We developed a methodology for the identification of FSTs using airborne laser scanning (ALS), and demonstrate its generality by applying it to forests from Boreal, Mediterranean and Atlantic biogeographical regions. First, hierarchal clustering analysis (HCA) was applied and clusters (FSTs) were determined in coniferous and deciduous forests using four forest structural variables obtained from forest inventory data – quadratic mean diameter (QMD), Gini coefficient (GC), basal area larger than mean (BALM) and density of stems (N) –. Then, classification and regression tree analysis (CART) were used to extract the empirical threshold values for discriminating those clusters. Based on the classification trees, GC and BALM were the most important variables in the identification of FSTs. Lower, medium and high values of GC and BALM characterize single storey FSTs, multi-layered FSTs and exponentially decreasing size distributions (reversed J), respectively. Within each of these main FST groups, we also identified young/mature and sparse/dense subtypes using QMD and N. Then we used similar structural predictors derived from ALS – maximum height (Max), L-coefficient of variation (Lcv), L-skewness (Lskew), and percentage of penetration (cover), – and a nearest neighbour method to predict the FSTs. We obtained a greater overall accuracy in deciduous forest (0.87) as compared to the coniferous forest (0.72). Our methodology proves the usefulness of ALS data for structural heterogeneity assessment of forests across biogeographical regions. Our simple two-tier approach to FST classification paves the way toward transnational assessments of forest structure across bioregions.

Most fragmentation research focuses on the effects of carving up old‐growth forests, but less is known about influences of habitat fragmentation on secondary succession in patches of regenerating forests. Working with forest dynamics on islands in a vast lake created by a hydroelectric dam in China (the Thousand Island Lake), we sampled 29 islands that were cleared of forest during dam construction in 1959 and then underwent succession. Measurements taken in 2009–2010 and 2014–2015 evaluated community assembly during succession, based on species diversity, functional traits and structural properties. Forests on small islands remained at relatively early stages of compositional succession: species richness was low, and communities were dominated by early‐successional species, with few animal‐dispersed and shade‐tolerant plants. However, these islands had accumulated similar above‐ground biomass density to that on larger islands, mostly in the stems of a fast‐growing pine species. Island size was the key driver of secondary succession in regenerating forest fragments, while isolation had comparatively little effect. Edge effects were important for species composition and functional composition. Synthesis. Habitat fragmentation resulting from the creation of an artificial lake affected community assembly, with fewer late‐successional species on smaller islands. Maintenance of large fragments is critical for the rapid succession of forests. Paradigms on habitat fragmentation effects drawn from old‐growth forest studies are unlikely to hold in regenerating forest fragments. Restoration activities should consider landscape patterns to accelerate secondary succession of regenerating forests.

Borneo contains some of the world's most biodiverse and carbon-dense tropical forest, but this 750 000 km2 island has lost 62 % of its old-growth forests within the last 40 years. Efforts to protect and restore the remaining forests of Borneo hinge on recognizing the ecosystem services they provide, including their ability to store and sequester carbon. Airborne laser scanning (ALS) is a remote sensing technology that allows forest structural properties to be captured in great detail across vast geographic areas. In recent years ALS has been integrated into statewide assessments of forest carbon in Neotropical and African regions, but not yet in Asia. For this to happen new regional models need to be developed for estimating carbon stocks from ALS in tropical Asia, as the forests of this region are structurally and compositionally distinct from those found elsewhere in the tropics. By combining ALS imagery with data from 173 permanent forest plots spanning the lowland rainforests of Sabah on the island of Borneo, we develop a simple yet general model for estimating forest carbon stocks using ALS-derived canopy height and canopy cover as input metrics. An advanced feature of this new model is the propagation of uncertainty in both ALS- and ground-based data, allowing uncertainty in hectare-scale estimates of carbon stocks to be quantified robustly. We show that the model effectively captures variation in aboveground carbon stocks across extreme disturbance gradients spanning tall dipterocarp forests and heavily logged regions and clearly outperforms existing ALS-based models calibrated for the tropics, as well as currently available satellite-derived products. Our model provides a simple, generalized and effective approach for mapping forest carbon stocks in Borneo and underpins ongoing efforts to safeguard and facilitate the restoration of its unique tropical forests.

Understanding the causes of variation in plant functional traits is a central issue in ecology, particularly in the context of global change. Spectroscopy is increasingly used for rapid and non-destructive estimation of foliar traits, but few studies have evaluated its accuracy when assessing phenotypic variation in multiple traits. Working with 24 chemical and physical leaf traits of six European tree species growing on strongly contrasting soil types (i.e. deep alluvium versus nearby shallow chalk), we asked (i) whether variability in leaf traits is greater between tree species or soil type; and (ii) whether field spectroscopy is effective at predicting intraspecific variation in leaf traits as well as interspecific differences. Analysis of variance showed that inter-specific differences in traits were generally much stronger than intraspecific differences related to soil type, accounting for 25% versus 5% of total trait variation, respectively. Structural traits, phenolic defences and pigments were barely affected by soil type. In contrast, foliar concentrations of rock-derived nutrients did vary: P and K concentration were lower on chalk than alluvial soils, while Ca, Mg, B, Mn and Zn concentrations were all higher, consistent with the findings of previous ecological studies. Foliar traits were predicted from 400-2500 nm reflectance spectra collected by field spectroscopy using partial least square regression, a method that is commonly employed in chemometrics. Pigments were best modelled using reflectance data from the visible region (400 - 700 nm), whilst all other traits were best modelled using reflectance data from the shortwave infrared region (1100 - 2500 nm) region. Spectroscopy delivered accurate predictions of species-level variation in traits. However, it was ineffective at detecting intraspecific variation in rock-derived nutrients (with the notable exception of P). The explanation for this failure is that rock-derived elements do not have absorption features in the 400-2500 nm region, and their estimation is indirect, relying on elemental concentrations co-varying with structural traits that do have absorption features in that spectral region (“constellation effects”). Since the structural traits did not vary with soil type, it was impossible for our regression models to predict intraspecific variation in rock-derived nutrients via constellation effects. This study demonstrates the value of spectroscopy for rapid, non-destructive estimation of foliar traits across species, but highlights problems with predicting intraspecific variation indirectly. We discuss the implications of these findings for mapping functional traits by airborne imaging spectroscopy.

Retrieval of forest biophysical properties using airborne LiDAR is known to differ between leaf-on and leaf-off states of deciduous trees, but much less is understood about the within-season effects of leafing phenology. Here, we compare two LiDAR surveys separated by just six weeks in spring, in order to assess whether LiDAR variables were influenced by canopy changes in Mediterranean mixed-oak woodlands at this time of year. Maximum and, to a slightly lesser extent, mean heights were consistently measured, whether for the evergreen cork oak (Quercus suber) or semi-deciduous Algerian oak (Q. canariensis) woodlands. Estimates of the standard deviation and skewness of height differed more strongly, especially for Algerian oaks which experienced considerable leaf expansion in the time period covered. Our demonstration of which variables are more or less affected by spring-time leafing phenology has important implications for analyses of both canopy and sub-canopy vegetation layers from LiDAR surveys.

Forest carbon stocks in rapidly developing tropical regions are highly heterogeneous, which challenges efforts to develop spatially-explicit conservation actions. In addition to field-based biodiversity information, mapping of carbon stocks can greatly accelerate the identification, protection and recovery of forests deemed to be of high conservation value (HCV). We combined airborne Light Detection and Ranging (LiDAR) with satellite imaging and other geospatial data to map forest aboveground carbon density at 30 m (0.09 ha) resolution throughout the Malaysian state of Sabah on the island of Borneo. We used the mapping results to assess how carbon stocks vary spatially based on forest use, deforestation, regrowth, and current forest protections. We found that unlogged, intact forests contain aboveground carbon densities averaging over 200 Mg C ha−1, with peaks of 500 Mg C ha−1. Critically, more than 40% of the highest carbon stock forests were discovered outside of areas designated for maximum protection. Previously logged forests have suppressed, but still high, carbon densities of 60–140 Mg C ha−1. Our mapped distributions of forest carbon stock suggest that the state of Sabah could double its total aboveground carbon storage if previously logged forests are allowed to recover in the future. Our results guide ongoing efforts to identify HCV forests and to determine new areas for forest protection in Borneo., Biological Conservation, 217, ISSN:0006-3207, ISSN:1873-2917

Brazil meets about 16% of its energy needs with bioethanol and other sugarcane products and leads the world in biofuel energy production. However, there are concerns over the environmental credentials of the sugarcane given the high emissions of nitrous oxide (N2O) − a potent greenhouse gas − associated with its production. We quantify N2O fluxes from a sugarcane experiment in the Brazilian Cerrado over an entire year, comparing the magnitude of response to different fertilizer treatments: mineral nitrogen (N), a liquid residue from bioethanol distillation known as vinasse (V) and a combination of both (NV), on irrigated and nonirrigated plots. We find that soil N2O fluxes in plots subjected to the NV treatment are, on average, at least three times higher than those in the other treatments, resulting in four times the emissions intensity per yield than when mineral N or vinasse were applied alone. We also find that irrigation has a positive effect on N2O emissions in the first two weeks after treatment addition and on annual sugarcane yield. Emission factors, defined as net N emissions as N2O as a percentage of the total mineral N added, vary from 0.05 to 4.6% according to treatment. Sugarcane production in the Cerrado is a significant source of N2O, due to the synergistic effect of mineral N fertilizer and vinasse. Using vinasse as the main fertilizer, or applying N and vinasse at widely spaced intervals are effective strategies for climate change mitigation, plant nutrition, waste management and cost-efficient production.

© 2018 IOP Publishing Ltd. Hyperspectral imaging is a cutting-edge type of remote sensing used for mapping vegetation properties, rock minerals and other materials. A major drawback of hyperspectral imaging devices is their intrinsic low spatial resolution. In this paper, we propose a method for increasing the spatial resolution of a hyperspectral image by fusing it with an image of higher spatial resolution that was obtained with a different imaging modality. This is accomplished by solving a variational problem in which the regularization functional is the directional total variation. To accommodate for possible mis-registrations between the two images, we consider a non-convex blind super-resolution problem where both a fused image and the corresponding convolution kernel are estimated. Using this approach, our model can realign the given images if needed. Our experimental results indicate that the non-convexity is negligible in practice and that reliable solutions can be computed using a variety of different optimization algorithms. Numerical results on real remote sensing data from plant sciences and urban monitoring show the potential of the proposed method and suggests that it is robust with respect to the regularization parameters, mis-registration and the shape of the kernel., MJE and C-BS acknowledge support from Leverhulme Trust project 'Breaking the non-convexity barrier', EPSRC grant 'EP/M00483X/1', EPSRC centre 'EP/N014588/1', the Cantab Capital Institute for the Mathematics of Information, and from CHiPS (Horizon 2020 RISE project grant). Moreover, C-BS is thankful for support from the Alan Turing Institute. DAC acknowledges the support of NERC (grant number NE/K016377/1). We are grateful to NERC's Airborne Research Facility and Data Analysis Node for conducting the airborne survey and pre-processing the environmental data collected from Alto Tajo, Spain (survey CAM11/03).

Southeast Asia is the epicentre of world palm oil production. Plantations in Malaysia have increased 150% in area within the last decade, mostly at the expense of tropical forests. Maps of the aboveground carbon density (ACD) of vegetation generated by remote sensing technologies, such as airborne LiDAR, are vital for quantifying the effects of land use change for greenhouse gas emissions, and many papers have developed methods for mapping forests. However, nobody has yet mapped oil palm ACD from LiDAR. The development of carbon prediction models would open doors to remote monitoring of plantations as part of efforts to make the industry more environmentally sustainable. This paper compares the performance of tree-centric and area-based approaches to mapping ACD in oil palm plantations. We find that an area-based approach gave more accurate estimates of carbon density than tree-centric methods and that the most accurate estimation model includes LiDAR measurements of top-of-canopy height and canopy cover. We show that tree crown segmentation is sensitive to crown density, resulting in less accurate tree density and ACD predictions, but argue that tree-centric approach can nevertheless be useful for monitoring purposes, providing a method to detect, extract and count oil palm trees automatically from images.

Forest canopy structure is strongly influenced by environmental factors and disturbance, and in turn influences key ecosystem processes including productivity, evapotranspiration and habitat availability. In tropical forests increasingly modified by human activities, the interplaying effects of environmental factors and disturbance legacies on forest canopy structure across landscapes are practically unexplored. We used high-fidelity airborne laser scanning (ALS) data to measure the canopy of old-growth and selectively logged peat swamp forest across a peat dome in Central Kalimantan, Indonesia, and quantified how canopy structure metrics varied with peat depth and under logging. Several million canopy gaps in different height cross-sections of the canopy were measured in 100 plots of 1 km2 spanning the peat dome, allowing us to describe canopy structure with seven metrics. Old-growth forest became shorter and had simpler vertical canopy profiles on deeper peat, consistently with previous work linking deep peat to stunted tree growth. Gap Size Frequency Distributions (GSFDs) indicated fewer and smaller canopy gaps on the deeper peat (i.e. the scaling exponent of pareto functions increased from 1.76 to 3.76 with peat depth). Areas subjected to concessionary logging until 2000, and informal logging since then, had the same canopy top height as old-growth forest, indicating the persistence of some large trees, but mean canopy height was significantly reduced; the total area of canopy gaps increased and the GSFD scaling exponent was reduced. Logging effects were most evident on the deepest peat, where nutrient depletion and waterlogged conditions restrain tree growth and recovery. A tight relationship exists between canopy structure and the peat deph gradient within the old-growth tropical peat swamp. This relationship breaks down after selective logging, with canopy structural recovery being modulated by environmental conditions.

The biodiversity–productivity relationship (BPR) is foundational to our understanding of the global extinction crisis and its impacts on ecosystem functioning. Understanding BPR is critical for the accurate valuation and effective conservation of biodiversity. Using ground-sourced data from 777,126 permanent plots, spanning 44 countries and most terrestrial biomes, we reveal a globally consistent positive concave-down BPR, whereby a continued biodiversity loss would result in an accelerating decline in forest productivity worldwide. The value of biodiversity in maintaining forest productivity—US$396–579 billion per year according to our estimation—is by itself over five times greater than the total cost of effective global conservation. This highlights the need for a worldwide re-assessment of biodiversity values, forest management strategies, and conservation priorities.

It has been suggested that diverse forests utilize canopy space more efficiently than species‐poor ones, as mixing species with complementary architectural and physiological traits allows trees to pack more densely. However, whether positive canopy packing–diversity relationships are a general feature of forests remains unclear. Using crown allometric data collected for 12 939 trees from permanent forest plots across Europe, we test (i) whether diversity promotes canopy packing across forest types and (ii) whether increased canopy packing occurs primarily through vertical stratification of tree crowns or as a result of intraspecific plasticity in crown morphology. We found that canopy packing efficiency increased markedly in response to species richness across a range of forest types and species combinations. Positive canopy packing–diversity relationships were primarily driven by the fact that trees growing in mixture had sizably larger crowns (38% on average) than those in monoculture. The ability of trees to plastically adapt the shape and size of their crowns in response to changes in local competitive environment is critical in allowing mixed‐species forests to optimize the use of canopy space. By promoting the development of denser and more structurally complex canopies, species mixing can strongly impact nutrient cycling and storage in forest ecosystems.

There is much current interest in using multisensor airborne remote sensing to monitor the structure and biodiversity of woodlands. This paper addresses the application of nonparametric (NP) image-registration techniques to precisely align images obtained from multisensor imaging, which is critical for the successful identification of individual trees using object recognition approaches. NP image registration, in particular, the technique of optimizing an objective function, containing similarity and regularization terms, provides a flexible approach for image registration. Here, we develop a NP registration approach, in which a normalized gradient field is used to quantify similarity, and curvature is used for regularization (NGF-Curv method). Using a survey of woodlands in southern Spain as an example, we show that NGF-Curv can be successful at fusing data sets when there is little prior knowledge about how the data sets are interrelated (i.e., in the absence of ground control points). The validity of NGF-Curv in airborne remote sensing is demonstrated by a series of experiments. We show that NGF-Curv is capable of aligning images precisely, making it a valuable component of algorithms designed to identify objects, such as trees, within multisensor data sets.

Characterising a species’ geographical extent is central to many conservation assessments, including those of the IUCN Red List of threatened species. The IUCN recommends that extent of occurrence (EOO) to be quantified by drawing a minimum convex polygon (MCP) around known or inferred presence localities. EOO calculated from verified specimens is commonly used in Red List assessments when other data are scarce, as is the case for many threatened plant species. Yet rarely do these estimates incorporate inferred localities from species distribution models (SDMs). A key impediment stems from uncertainty about how SDM predictions relate to EOO. Here we address this issue by comparing the EOOs estimated from specimen localities with EOOs derived from SDMs for plant species occurring in Costa Rica and Panama. We first analyse 20 plant species, with well-known and well-sampled distributions, and train SDMs to subsamples of the data and assess how well the SDM-derived MCPs predict both the MCPs of the subsamples and the MCPs of the complete dataset. We find that when sample sizes are small (5 or 10 samples) the SDM-derived MCPs are actually closer to the complete dataset than to the MCPs of the subsamples, both in terms of EOO and geographically. This occurs when using a probability threshold based on maximum geographical similarity between the SDM-derived MCP and the subsample MCP; other threshold methods performed less well. For the species with less well-known distributions, the SDM-derived EOOs correlate strongly with, but tend to be larger than, EOOs estimated by point data. This implies that a SDM-derived EOO may be more representative of the full EOO than that drawn around known localities. Our findings reveal situations in which SDMs provide useful information that complements the IUCN Red Listing process.

Remote sensing of individual tree species has many applications in resource management, biodiversity assessment, and conservation. Airborne remote sensing using light detection and ranging (LiDAR) and hyperspectral sensors has been used extensively to extract biophysical traits of vegetation and to detect species. However, its application for individual tree mapping remains limited due to the technical challenges of precise coalignment of images acquired from different sensors and accurately delineating individual tree crowns (ITCs). In this study, we developed a generic workflow to map tree species at ITC level from hyperspectral imagery and LiDAR data using a combination of well established and recently developed techniques. The workflow uses a nonparametric image registration approach to coalign images, a multiclass normalized graph cut method for ITC delineation, robust principal component analysis for feature extraction, and support vector machine for species classification. This workflow allows us to automatically map tree species at both pixel- and ITC-level. Experimental tests of the technique were conducted using ground data collected from a fully mapped temperate woodland in the UK. The overall accuracy of pixel-level classification was 91%, while that of ITC-level classification was 61%. The test results demonstrate the effectiveness of the approach, and in particular the use of robust principal component analysis to prune the hyperspectral dataset and reveal subtle difference among species.

The evaluation of accuracy is essential for assuring the reliability of ecological models. Usually, the accuracy of above-ground biomass (AGB) predictions obtained from remote sensing is assessed by the mean differences (MD), the root mean squared differences (RMSD), and the coefficient of determination (R^2) between observed and predicted values. In this article we propose a more thorough analysis of accuracy, including a hypothesis test to evaluate the agreement between observed and predicted values, and an assessment of the degree of overfitting to the sample employed for model training. Using the estimation of forest AGB from LIDAR and spectral sensors as a case study, we compared alternative prediction and variable selection methods using several statistical measures to evaluate their accuracy. We showed that the hypothesis tests provide an objective method to infer the statistical significance of agreement. We also observed that overfitting can be assessed by comparing the inflation in residual sums of squares experienced when carrying out a cross-validation. Our results suggest that this method may be more effective than analysing the deflation in R^2. We proved that overfitting needs to be specifically addressed since, in light of MD, RMSD and R^2 alone, predictions may apparently seem reliable even in clearly unrealistic circumstances, for instance when including too many predictor variables. Moreover, Theil’s partial inequality coefficients, which are employed to resolve the proportions of the total errors due to the unexplained variance, the slope and the bias, may become useful to detect averaging effects common in remote sensing predictions of AGB. We concluded that statistical measures of accuracy, precision and agreement are necessary but insufficient for model evaluation. We therefore advocate for incorporating evaluation measures specifically devoted to testing observed-versus-predicted fit, and to assessing the degree of overfitting.

© 2017 by the authors. We apply a spatially-implicit, allometry-based modelling approach to predict stem diameter distributions (SDDs) from low density airborne LiDAR data in a heterogeneous, temperate forest in Ontario, Canada. Using a recently published algorithm that relates the density, size, and species of individual trees to the height distribution of first returns, we estimated parameters that succinctly describe SDDs that are most consistent with each 0.25-ha LiDAR tile across a 30,000 ha forest landscape. Tests with independent validation plots showed that the diameter distribution of stems was predicted with reasonable accuracy in most cases (half of validation plots had R2 ≥ 0.75, and another 23% had 0.5 ≤ R2 < 0.75). The predicted frequency of larger stems was much better than that of small stems (8 ≤ x < 11 cm diameter), particularly small conifers. We used the predicted SDDs to calculate aboveground carbon density (ACD; RMSE = 21.4 Mg C/ha), quadratic mean diameter (RMSE = 3.64 cm), basal area (RMSE = 6.99 m2/ha) and stem number (RMSE = 272 stems/ha). The accuracy of our predictions compared favorably with previous studies that have generally been undertaken in simpler conifer-dominated forest types. We demonstrate the utility of our results to spatial forest management planning by mapping SDDs, the proportion of broadleaves, and ACD at a 0.25 ha resolution.

© 2017, Canadian Science Publishing. All rights reserved. Estimation of the Gini coefficient (GC) of tree sizes using airborne laser scanning (ALS) can provide maps of forest structure across the landscape, which can support sustainable forest management. A challenge arise s in determining the optimal spatial resolution that maximizes the stability and precision of GC estimates, which in turn depends on stand density or ALS scan density. By subsampling different plot sizes within large field plots, we evaluated the optimal spatial resolution by observing changes in GC estimation and in its correlation with ALS metrics. We found that plot size had greater effects than either stand density or ALS scan density on the relationship between GC and ALS metrics. Uncertainty in GC estimates fell as plot size increased. Correlation with ALS metrics showed convex curves with maxima at 250–450m 2 , which thus was considered the optimal plot size and, consequently, the optimal spatial resolution. By thinning the density of the ALS point cloud, we deduced that at least 3 points·m −2 were needed for reliable GC estimates. Many nationwide ALS scan densities are sparser than this, so may be unreliable for GC estimation. Ours is a simple approach for evaluating the optimal spatial resolution in remote sensing estimation of any forest attribute.

Understanding the causes of variation in plant functional traits is a central issue in ecology, particularly in the context of global change. Spectroscopy is increasingly used for rapid and non-destructive estimation of foliar traits, but few studies have evaluated its accuracy when assessing phenotypic variation in multiple traits. Working with 24 chemical and physical leaf traits of six European tree species growing on strongly contrasting soil types (i.e. deep alluvium versus nearby shallow chalk), we asked (i) whether variability in leaf traits is greater between tree species or soil type; and (ii) whether field spectroscopy is effective at predicting intraspecific variation in leaf traits as well as interspecific differences. Analysis of variance showed that inter-specific differences in traits were generally much stronger than intraspecific differences related to soil type, accounting for 25% versus 5% of total trait variation, respectively. Structural traits, phenolic defences and pigments were barely affected by soil type. In contrast, foliar concentrations of rock-derived nutrients did vary: P and K concentration were lower on chalk than alluvial soils, while Ca, Mg, B, Mn and Zn concentrations were all higher, consistent with the findings of previous ecological studies. Foliar traits were predicted from 400-2500 nm reflectance spectra collected by field spectroscopy using partial least square regression, a method that is commonly employed in chemometrics. Pigments were best modelled using reflectance data from the visible region (400 - 700 nm), whilst all other traits were best modelled using reflectance data from the shortwave infrared region (1100 - 2500 nm) region. Spectroscopy delivered accurate predictions of species-level variation in traits. However, it was ineffective at detecting intraspecific variation in rock-derived nutrients (with the notable exception of P). The explanation for this failure is that rock-derived elements do not have absorption features in the 400-2500 nm region, and their estimation is indirect, relying on elemental concentrations co-varying with structural traits that do have absorption features in that spectral region (“constellation effects”). Since the structural traits did not vary with soil type, it was impossible for our regression models to predict intraspecific variation in rock-derived nutrients via constellation effects. This study demonstrates the value of spectroscopy for rapid, non-destructive estimation of foliar traits across species, but highlights problems with predicting intraspecific variation indirectly. We discuss the implications of these findings for mapping functional traits by airborne imaging spectroscopy.

© 2020 The Authors. Biotropica published by Wiley Periodicals LLC on behalf of Association for Tropical Biology and Conservation The factors that limit the maximum height of trees, whether ecophysiological or mechanical, are the subject of longstanding debate. Here, we examine the role of mechanical stability in limiting tree height and focus on trees from the tallest tropical forests on Earth, in Sabah, Malaysian Borneo, including the recently discovered tallest tropical tree, a 100.8 m Shorea faguetiana named Menara. We use terrestrial laser scans, in situ strain gauge data and finite element simulations, to map the architecture of tall tropical trees and monitor their response to wind loading. We demonstrate that a tree's risk of breaking due to gravity or self-weight decreases with tree height and is much more strongly affected by tree architecture than by material properties. In contrast, wind damage risk increases with tree height despite the larger diameters of tall trees, resulting in a U-shaped curve of mechanical risk with tree height. Our results suggest that the relative rarity of extreme wind speeds in north Borneo may be the reason it is home to the tallest trees in the tropics. Abstract in MALAY is available with online material.

Agricultural expansion is the primary driver of ecological degradation across the tropics. Set-asides—uncultivated parts of agricultural landscapes, often on steep slopes and alongside rivers—may alleviate environmental impacts but can reduce the area cultivated. Here we model an approach to configuring set-asides aimed at optimizing ecological outcomes (biodiversity, above-ground carbon storage and nutrient cycling) without reducing net cultivation area. We compare set-asides in an oil palm landscape where all plantations adopt the same configuration (‘uniform’ approach) with a scenario where there can be variation in configuration among plantations (‘variable’ approach). We find that all set-aside configurations support substantial ecological values but that the best strategies involve set-asides, particularly alongside rivers, that are spatially targeted and variable among plantations. This ‘variable’ approach can increase ecological outcomes twofold over the ‘uniform’ approach without reducing net cultivation area. Our findings underscore the potential importance of well-planned set-asides for enhancing agricultural sustainability.

[Abstract] Here we provide the ‘Global Spectrum of Plant Form and Function Dataset’, containing species mean values for six vascular plant traits. Together, these traits –plant height, stem specific density, leaf area, leaf mass per area, leaf nitrogen content per dry mass, and diaspore (seed or spore) mass – define the primary axes of variation in plant form and function. The dataset is based on ca. 1 million trait records received via the TRY database (representing ca. 2,500 original publications) and additional unpublished data. It provides 92,159 species mean values for the six traits, covering 46,047 species. The data are complemented by higher-level taxonomic classification and six categorical traits (woodiness, growth form, succulence, adaptation to terrestrial or aquatic habitats, nutrition type and leaf type). Data quality management is based on a probabilistic approach combined with comprehensive validation against expert knowledge and external information. Intense data acquisition and thorough quality control produced the largest and, to our knowledge, most accurate compilation of empirically observed vascular plant species mean traits to date. The study has been supported by the TRY initiative on plant traits (https://www.try-db.org). TRY is an initiative of the Max Planck Institute for Biogeochemistry, bioDISCOVERY/Future Earth (ICSU), the German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig and Núcleo DiverSus (CONICET- Universidad Nacional de Córdoba, Argentina). The Global Spectrum of Plant Form and Function study has been supported by the European BACI project (Towards a Biosphere Atmosphere change Index, EU grant ID 640176), and grants to SD by FONCyT, CONICET, Universidad Nacional de Córdoba, the Inter-American Institute for Global Change Research, and The Newton Fund (NERC UK – CONICET ARG). VO thanks RSF (#19-14-00038p). Open Access funding enabled and organized by Projekt DEAL

Canopy gaps are openings in the forest canopy resulting from branch fall and tree mortality events. The geographical distribution of large canopy gaps may reflect underlying variation in mortality and growth processes. However, a lack of data at the appropriate scale has limited our ability to study this relationship until now. We detected canopy gaps using a unique LiDAR dataset consisting of 650 transects randomly distributed across 2500 km(2) of the Brazilian Amazon. We characterized the size distribution of canopy gaps using a power law and we explore the variation in the exponent, alpha. We evaluated how the alpha varies across the Amazon, in response to disturbance by humans and natural environmental processes that influence tree mortality rates. We observed that South-eastern forests contained a higher proportion of large gaps than North-western, which is consistent with recent work showing greater tree mortality rates in the Southeast than the Northwest. Regions characterized by strong wind gust speeds, frequent lightning and greater water shortage also had a high proportion of large gaps, indicating that geographical variation in alpha is a reflection of underlying disturbance processes. Forests on fertile soils were also found to contain a high proportion of large gaps, in part because trees grow tall on these sites and create large gaps when they fall; thus, canopy gap analysis picked up differences in growth as well as mortality processes. Finally, we found that human-modified forests had a higher proportion of large gaps than intact forests, as we would expect given that these forests have been disturbed. Synthesis. The proportion of large gaps in the forest canopy varied substantially over the Brazilian Amazon. We have shown that the trends can be explained by geographical variation in disturbance and growth. The frequency of extreme weather events is predicted to increase under climate change, and changes could lead to greater forest disturbance, which should be detectable as an increased proportion of large gaps in intact forests.

© 2019 The Authors. Journal of Ecology © 2019 British Ecological Society The infrequent and unpredictable nature of earthquakes means that their landslide-generated impacts on forests are rarely investigated. In montane forests, landslides are the main cause of tree death and injury during earthquakes. Landslides range from soil movements that uproot and bury trees over extensive areas to rock falls that strike individual trees. We examined unexplored relationships between tree survival and distance from an epicentre, soil-available phosphorus (P) as an indicator of soil development and tree diameter. We expected decreased tree growth in damaged forests because of tree injury. We used a plot network, established in 1974 and resurveyed regularly ever since, to quantify survival and growth responses 6–30 km from the epicentre of a 1994 earthquake in New Zealand's Southern Alps. Our Bayesian analysis used 8,518 trees from 250 plots that representatively sampled a naturally monospecific Nothofagus forest. As the time-scales over which responses could emerge were unknown, we compared relationships for a pre-earthquake period with 0–5 years post-earthquake, and with 5+ years post-earthquake. Not all plots were affected by the earthquake. We found that 0–5 years post-earthquake survival increased logarithmically with distance from the epicentre with lowered survival up to 20 km from the epicentre. Survival was low on plots with high soil-available P. An inverted U-shaped relationship between survival and diameter pre-earthquake was not found 0–5 years post-earthquake. This was because of surprisingly high survival by large trees. The earthquake most often suppressed 0–5 years post-earthquake growth up to 15 km from the epicentre, but this was only apparent after accounting for more general growth differences among periods. The positive relationship between growth and soil-available P pre-earthquake and 5+ years post-earthquake reflected enhanced growth on young soils. This contrasted with a negative effect of soil-available P on growth 0–5 years post-earthquake. Synthesis. Soil-available P, tree diameter and distance from the epicentre independently determined how tree survival and growth responded to an earthquake. The impacts on survival and growth largely occurred 0–5 years post-earthquake and suggests a level of resilience in mountain beech forests.

An enduring challenge in ecology is to characterise and understand species richness patterns in tropical regions. Species richness maps produced by stacking species distribution model (SDM) range maps could prove useful in this regard, but little attention has been given to this approach. Here we generate a species richness map by stacking the ranges of 86 Mesoamerican fern species modelled by MaxEnt from field data collected by the Sampled Red List Index for Plants project. Predicted species richness showed a hump-backed relationship with elevation, peaking at mid-elevation (1800–2000 m). A remarkably similar pattern was observed in a field survey conducted to validate the approach. Predicted species richness was also low in sites with high water deficits, as previously shown in the fern literature. Beta-diversity in the lowlands was greatest between sites with strongly contrasting water deficits, further emphasising the importance of this environmental variable. The stacked SDM approach was thus able to reproduce broad biogeographical patterns of species richness, despite many of the fern species being represented by fewer than 20 samples. Keywords: MaxEnt, Mesoamerican ferns, Sampled red list index (SRLI), Species distribution models (SDMs)

Background Forests are a key component of the global carbon cycle, and research is needed into the effects of human-driven and natural processes on their carbon pools. Airborne laser scanning (ALS) produces detailed 3D maps of forest canopy structure from which aboveground carbon density can be estimated. Working with a ALS dataset collected over the 8049-km2 Wellington Region of New Zealand we create maps of indigenous forest carbon and evaluate the influence of wind by examining how carbon storage varies with aspect. Storms flowing from the west are a common cause of disturbance in this region, and we hypothesised that west-facing forests exposed to these winds would be shorter than those in sheltered east-facing sites. Methods The aboveground carbon density of 31 forest inventory plots located within the ALS survey region were used to develop estimation models relating carbon density to ALS information. Power-law models using rasters of top-of-the-canopy height were compared with models using tree-level information extracted from the ALS dataset. A forest carbon map with spatial resolution of 25 m was generated from ALS maps of forest height and the estimation models. The map was used to evaluate the influences of wind on forests. Results Power-law models were slightly less accurate than tree-centric models (RMSE 35% vs 32%) but were selected for map generation for computational efficiency. The carbon map comprised 4.5 million natural forest pixels within which canopy height had been measured by ALS, providing an unprecedented dataset with which to examine drivers of carbon density. Forests facing in the direction of westerly storms stored less carbon, as hypothesised. They had much greater above-ground carbon density for a given height than any of 14 tropical forests previously analysed by the same approach, and had exceptionally high basal areas for their height. We speculate that strong winds have kept forests short without impeding basal area growth. Conclusion Simple estimation models based on top-of-the canopy height are almost as accurate as state-of-the-art tree-centric approaches, which require more computing power. High-resolution carbon maps produced by ALS provide powerful datasets for evaluating the environmental drivers of forest structure, such as wind.