Search

Coral reef research and management efforts can be improved when supported by reef maps providing local-scale details across global extents. However, such maps are difficult to generate due to the broad geographic range of coral reefs, the complexities of relating satellite imagery to geomorphic or ecological realities, and other challenges. However, reef extent maps are one of the most commonly used and most valuable data products from the perspective of reef scientists and managers. Here, we used convolutional neural networks to generate a globally consistent coral reef probability map—a probabilistic estimate of the geospatial extent of reef ecosystems—to facilitate scientific, conservation, and management efforts. We combined a global mosaic of high spatial resolution Planet Dove satellite imagery with regional Millennium Coral Reef Mapping Project reef extents to build training, validation, and application datasets. These datasets trained our reef extent prediction model, a neural network with a dense-unet architecture followed by a random forest classifier, which was used to produce a global coral reef probability map. Based on this probability map, we generated a global coral reef extent map from a 60% threshold of reef probability (reef: probability ≥ 60%, non-reef: probability

High-resolution benthic habitat data fill an important knowledge gap for many areas of the world and are essential for strategic marine conservation planning and implementing effective resource management. Many countries lack the resources and capacity to create these products, which has hindered the development of accurate ecological baselines for assessing protection needs for coastal and marine habitats and monitoring change to guide adaptive management actions. The PlanetScope (PS) Dove Classic SmallSat constellation delivers high-resolution imagery (4 m) and near-daily global coverage that facilitates the compilation of a cloud-free and optimal water column image composite of the Caribbean’s nearshore environment. These data were used to develop a first-of-its-kind regional thirteen-class benthic habitat map to 30 m water depth using an object-based image analysis (OBIA) approach. A total of 203,676 km2 of shallow benthic habitat across the Insular Caribbean was mapped, representing 5% coral reef, 43% seagrass, 15% hardbottom, and 37% other habitats. Results from a combined major class accuracy assessment yielded an overall accuracy of 80% with a standard error of less than 1% yielding a confidence interval of 78–82%. Of the total area mapped, 15% of these habitats (31,311.7 km2) are within a marine protected or managed area. This information provides a baseline of ecological data for developing and executing more strategic conservation actions, including implementing more effective marine spatial plans, prioritizing and improving marine protected area design, monitoring condition and change for post-storm damage assessments, and providing more accurate habitat data for ecosystem service models.

Coral reefs are among the most diverse and iconic ecosystems on Earth, but a range of anthropogenic pressures are threatening their persistence. Owing to their remoteness, broad spatial coverage and cross‐jurisdictional locations, there are no high‐resolution remotely sensed maps available at the global scale. Here we present a framework that is capable of mapping coral reef habitats from individual reefs (~200 km²) to entire barrier reef systems (200 000 km²) and across vast ocean extents (>6 000 000 km²). This is the first time this has been demonstrated using a consistent and transparent remote sensing mapping framework. The ten maps that we present achieved good accuracy (78% mean overall accuracy) from multiple input image datasets and training data sources, and our framework was shown to be adaptable to either benthic or geomorphic reef features and across diverse coral reef environments. These new generation high‐resolution map data will be useful for supporting ecosystem risk assessments, detecting change in ecosystem dynamics and targeting efforts to monitor local‐scale changes in coral cover and reef health.

Remote sensing data have increasingly been employed in combination with field plot data to estimate aboveground carbon (C) stocks across heterogeneous forested landscapes around the world. The Forest Inventory and Analysis (FIA) program of the US Forest Service offers a gridded network of field plots which potentially can be linked to airborne Light Detection and Ranging (LiDAR) data to estimate forest aboveground carbon density (ACD; units of Mg C ha−1). Here we utilized FIA plot and airborne LiDAR data sets collected across two contrasting landscapes known as Laupahoehoe and Pu‘u Wa‘awa‘a on Hawai’i Island to explore strengths and weaknesses of linking those two data sets to estimate ACD. We varied FIA plot sample designs with respect to sampling density (i.e., the number of plots across landscape) and intensity (i.e., the structural detail within inventory plots) to test the capability of the mapping approach. Results indicated that Laupahoehoe and Pu‘u Wa‘awa‘a landscapes supported an estimated 545 Gg C and 157 Gg C aboveground, respectively, and mean ACD values of the wet windward Laupahoehoe landscape (109 Mg ha−1) were an order of magnitude greater than those of the leeward dry Pu‘u Wa‘awa‘a landscape (9.7 Mg ha−1). Patterns of ACD were largely determined by combined factors of precipitation, lava substrate, prior land use, and presence of non-native, often invasive, species. Results demonstrated the relative importance of sample plot density over sample plot intensity, and showed that FIA inventory plots, even at their lowest sample intensity design, can be linked with LiDAR data to accurately estimate ACD across spatially heterogeneous landscapes. We also developed and applied a straightforward, statistically-robust approach to provide error estimates for the 100 million pixels that characterize the Laupahoehoe and Pu‘u Wa‘awa‘a landscapes as well as for any sub-units of those landscapes. We contend that augmenting existing FIA forest plot data with airborne LiDAR coverage, even if that requires an increase in plot density somewhat above the FIA standard 1X or 2X approaches, is a feasible, cost-effective, scientifically sound approach from which to obtain accurate landscape- to regional-scale ACD measures across the extensive and heterogeneous forests of the United States.

Remote imaging spectroscopy from 400 to 800 nm can use benthic reflectance signatures to map the composition and condition of shallow water ecosystems. We present a novel probabilistic approach to jointly estimate the seafloor reflectance and water properties while flexibly incorporating varied domain knowledge and in situ measurements. The inversion transforms remote radiance data with an atmospheric correction followed by a water column correction. Benthic reflectance and water optical properties are both represented by linear mixtures of endmember spectra. We combine remote measurements, prior knowledge and field data using a flexible Bayesian optimal estimation, solving for the Maximum A Posteriori (MAP) combination of water column properties, seafloor reflectance, and depth. We then demonstrate performance in controlled simulations and in overflights of a coral reef in Hawaii with coincident in situ measurements. The measurement approach helps lay a foundation for wide-area airborne mapping of the condition of threatened coastal ecosystems such as coral reefs.

Global shallow water bathymetry maps offer critical information to inform activities such as scientific research, environment protection, and marine transportation. Methods that employ satellite-based bathymetric modeling provide an alternative to conventional shipborne measurements, offering high spatial resolution combined with extensive coverage. We developed an automated bathymetry mapping approach based on the Sentinel-2 surface reflectance dataset in Google Earth Engine. We created a new method for generating a clean-water mosaic and a tailored automatic bathymetric estimation algorithm. We then evaluated the performance of the models at six globally diverse sites (Heron Island, Australia; West Coast of Hawaiʻi Island, Hawaiʻi; Saona Island, Dominican Republic; Punta Cana, Dominican Republic; St. Croix, United States Virgin Islands; and The Grenadines) using 113,520 field bathymetry sampling points. Our approach derived accurate bathymetry maps in shallow waters, with Root Mean Square Error (RMSE) values ranging from 1.2 to 1.9 m. This automatic, efficient, and robust method was applied to map shallow water bathymetry at the global scale, especially in areas which have high biodiversity (i.e., coral reefs).

Average responses of forest foliar traits to elevation are well understood, but far less is known about trait distributional responses to elevation at multiple ecological scales. This limits our understanding of the ecological scales at which trait variation occurs in response to environmental drivers and change. We analyzed and compared multiple canopy foliar trait distributions using field sampling and airborne imaging spectroscopy along an Andes-to-Amazon elevation gradient. Field-estimated traits were generated from three community-weighting methods, and remotely sensed estimates of traits were made at three scales defined by sampling grain size and ecological extent. Field and remote sensing approaches revealed increases in average leaf mass per unit area (LMA), water, nonstructural carbohydrates (NSCs) and polyphenols with increasing elevation. Foliar nutrients and photosynthetic pigments displayed little to no elevation trend. Sample weighting approaches had little impact on field-estimated trait responses to elevation. Plot representativeness of trait distributions at landscape scales decreased with increasing elevation. Remote sensing indicated elevation-dependent increases in trait variance and distributional skew. Multiscale invariance of LMA, leaf water and NSC mark these traits as candidates for tracking forest responses to changing climate. Trait-based ecological studies can be greatly enhanced with multiscale studies made possible by imaging spectroscopy. Fil: Asner, Gregory P.. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Martin, Roberta E.. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Anderson, Christopher Brian. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina Fil: Kryston, Katherine. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Vaughn, Nicholas. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Knapp, David E.. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Bentley, Lisa Patrick. University of Oxford; Reino Unido Fil: Shenkin, Alexander. University of Oxford; Reino Unido Fil: Salinas, Norma. University of Oxford; Reino Unido. Pontificia Universidad Católica de Perú; Perú Fil: Sinca, Felipe. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Tupayachi, Raul. Carnegie Institution for Science. Department of Global Ecology; Estados Unidos Fil: Quispe Huaypar, Katherine. Universidad Nacional de San Antonio Abad del Cusco; Perú Fil: Montoya Pillco, Milenka. Universidad Nacional de San Antonio Abad del Cusco; Perú Fil: Ccori Álvarez, Flor Delis. Universidad Nacional de San Antonio Abad del Cusco; Perú Fil: Díaz, Sandra Myrna. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba. Instituto Multidisciplinario de Biología Vegetal. Universidad Nacional de Córdoba. Facultad de Ciencias Exactas Físicas y Naturales. Instituto Multidisciplinario de Biología Vegetal; Argentina Fil: Enquist, Brian J.. Arizona State University; Estados Unidos Fil: Malhi, Yadvinder. University of Oxford; Reino Unido

Context Many arboreal mammals in Neotropical forests are important seed dispersers that influence the spatial patterns of tree regeneration via their movement patterns, which in turn are determined by the canopy structure of the forest itself. However, the relationship between arboreal mammal movement and canopy structure is poorly understood, due in large part to the complexity of quantifying arboreal habitat structure. Objectives We relate detailed movement trajectories of three sympatric primate species to attributes of canopy structure derived from airborne light detection and ranging (LiDAR) in order to understand the role of structure in arboreal movement in the tropical moist forest of Barro Colorado Island, Panama. Methods We used high-resolution LiDAR to quantify three-dimensional attributes of the forest canopy of the entire island, high-resolution GPS tracking to map the movement patterns of the monkey species, and step selection functions to relate movement decisions to canopy attributes. Results We found that movement decisions were correlated with canopy height and distance to gaps, which indicate forest maturity and lateral connectivity, in all three species. In the two faster-moving species, step selection was also correlated with the thickness of the crown layer and the density of vegetation within the crown. Conclusions The correlations detected are fully in line with known differences in the locomotor adaptations and movement strategies of the study species, and directly reflect maximization of energetic efficiency and ability to escape from predators. Quantification of step selection in relation to structure thus provides insight into the ways in which arboreal animals use their environment.

We used imaging spectroscopy to perform a top-down mesoscale analysis of tropical tree species richness across a bioclimatic gradient in Panama. The expressed precipitation gradient from the wet Caribbean side to the dry Pacific side makes Panama an excellent study area for performing a mesoscale assessment of climate effects on tropical tree species richness. Spatial patterns in local spectral variability (expressed as the coefficient of variation) and spectral similarity (expressed as the spectral similarity index) were used as proxies for species area curves and species distance decay curves. Our analysis revealed significant spectral changes along the precipitation gradient. Highest spectral diversity was observed for moist forest sites while lowest diversity was observed for the driest forest sites. Most of the spectral variation came from changes in the visible (VIS) and shortwave-infrared (SWIR) reflectance. Variation in the VIS was significantly higher for the dry compared to the moist and wet forests, while the opposite was true for the NIR and SWIR reflectance. Our spectral mesoscale analysis extends previous results suggesting that niche differentiation with respect to soil water availability is a direct determinant of both local- and regional-scale distributions of tropical trees. A next step would be to test the accuracy and scalability of our results with lower spatial resolution spectrometer data, simulating the observing conditions that will be achieved with future satellite missions such as the European Union's EnMap and NASA's HyspIRI missions.

Tropical forest functional diversity, which is a measure of the diversity of organismal interactions with the environment, is poorly understood despite its importance for linking evolutionary biology to ecosystem biogeochemistry. Functional diversity is reflected in functional traits such as the concentrations of different compounds in leaves or the density of leaf mass, which are related to plant activities such as plant defence, nutrient cycling, or growth. In the Amazonian lowlands, river movement and microtopography control nutrient mobility, which may influence functional trait distributions. Here we use airborne laser-guided imaging spectroscopy to develop maps of 16 forest canopy traits, throughout four large landscapes that harbour three common forest community types on the Madre de Dios and Tambopata rivers in southwestern Amazonia. Our maps, which are based on quantitative chemometric analysis of forest canopies with visible-to-near infrared (400–2,500 nm) spectroscopy, reveal substantial variation in canopy traits and their distributions within and among forested landscapes. Forest canopy trait distributions are arranged in a nested pattern, with location along rivers controlling trait variation between different landscapes, and microtopography controlling trait variation within landscapes. We suggest that processes of nutrient deposition and depletion drive increasing phosphorus limitation, and a corresponding increase in plant defence, in an eastward direction from the base of the Andes into the Amazon Basin. The controls on plant functional diversity are unclear. Analysis of spectral data from the tree canopy in the Amazonian lowlands implies that plant functional traits are influenced by nutrient supply, which in turn varies with topography.

Spatial and temporal information on plant functional traits are lacking in ecology, which limits our understanding of how plant communities and ecosystems are changing. This problem is acute in remote tropical regions, where information on plant functional traits is difficult to ascertain. We used Carnegie Airborne Observatory visible-to-shortwave infrared (VSWIR) imaging spectroscopy with light detection and ranging (LiDAR) to assess the foliar traits of Amazonian and Andean tropical forest canopies. We calibrated and validated the retrieval of 15 canopy foliar chemicals and leaf mass per area (LMA) across a network of 79 1-hectare field plots using a new VSWIR-LiDAR fusion approach designed to accommodate the enormous scale mismatch between field and remote sensing studies. The results indicate that sparse and highly variable field sampling can be integrated with VSWIR-LiDAR data to yield demonstrably accurate estimates of canopy foliar chemical traits. This new airborne approach addresses the inherent limitations and sampling biases associated with field-based studies of forest functional traits, particularly in structurally and floristically complex tropical canopies.

Landsat and Shuttle Radar Topography Mission (SRTM) imagery have recently been used to identify broad-scale floristic units in Neotropical rain forests, corresponding to geological formations and their edaphic properties. Little is known about the structural and functional variation between these floristic units, however, and Landsat and SRTM data lack the spectral and spatial resolution needed to provide this information. Imaging spectroscopy and LiDAR (Light Detection and Ranging) have been used to measure canopy structure and function in a variety of ecosystems, but the ability of these technologies to measure differences between compositionally-distinct but otherwise uniform tropical forest types remains unknown. We combined 16 tree inventories from central Panama with imaging spectroscopy and LiDAR elevation data from the Carnegie Airborne Observatory to test our ability to identify patterns in plant species composition, and to measure the spectral and structural differences between adjacent closed-canopy tropical forest types. We found that variations in spectroscopic imagery and LiDAR data were strong predictors of spatial turnover in plant species composition. We also found that these compositional, chemical, and structural patterns corresponded to underlying geological formations and their geomorphological properties. We conclude that imaging spectroscopy and LiDAR data can be used to interpret patterns identified in lower resolution sensors, to provide new information on forest function and structure, and to identify underlying determinants of these patterns.

Elevation gradients provide opportunities to explore environmental controls on forest structure and functioning. We used airborne imaging spectroscopy and lidar (light detection and ranging) to quantify changes in three-dimensional forest structure and canopy functional traits in twenty 25 ha landscapes distributed along a 3300 m elevation gradient from lowland Amazonia to treeline in the Peruvian Andes. Elevation was positively correlated with lidar-estimated canopy gap density and understory vegetation cover, and negatively related to canopy height and the vertical partitioning of vegetation in canopies. Increases in canopy gap density were tightly linked to increases in understory plant cover, and larger gaps (20–200 m2) produced 25–30 times the response in understory cover than did smaller gaps (< 5 m2). Vegetation NDVI and photosynthetic fractional cover decreased, while exposed non-photosynthetic vegetation and bare soil increased, with elevation. Scaling of gap size to gap frequency (λ) was, however, nearly constant along the elevation gradient. When combined with other canopy structural and functional trait information, this suggests near-constant canopy turnover rates from the lowlands to treeline, which occurs independent of decreasing biomass or productivity with increasing elevation. Our results provide the first landscape-scale quantification of forest structure and canopy functional traits with changing elevation, thereby improving our understanding of disturbance, demography and ecosystem processes in the Andes-to-Amazon corridor.

Measurement of vegetation biomass accumulation is critical for ecosystem assessment and monitoring, but doing so typically involves extensive field data collection that yields relatively crude structural outputs, e.g., plot- or site-level metrics. This study assessed the utility of airborne light detection and ranging (lidar) waveform features to explain structural and biomass variation in a savanna ecosystem across a land-use gradient. The ability of aboveground waveform lidar features to model field-based woody and herbaceous biomass measurements was evaluated statistically by regression models using forward variable selection. Waveform features explained 76% of the variation in woody biomass in a regulated communal land use area (RMSE = 29.0 kg). The waveform features were also correlated to herbaceous measurements in the same land-use area, with increased correlations at higher biomass levels. These results indicate that small-footprint waveform lidar data potentially can be used as a single modality to describe heterogeneous woody cover in a savanna environment; however, further research is warranted during the full growing season to fully evaluate its performance.

Airborne spectroscopy for forest traitsThe development of conservation priorities in the tropics is often hampered by the sparseness of ground data on biological diversity and the relative crudeness of larger-scale remote sensing data. Asneret al.developed airborne instrumentation to make large-scale maps of forest functional diversity across 72 million hectares of the Peruvian Andes and Amazon basin (see the Perspective by Kapos). They generated a suite of forest canopy functional trait maps from laser-guided imaging spectroscopy and used them to define distinct forest functional classes. These were then compared with government deforestation and land allocation data, which enabled an analysis of conservation threats and opportunities across the region.Science, this issue p.385; see also p.347

Leaf economics spectrum (LES) theory suggests a universal trade-off between resource acquisition and storage strategies in plants, expressed in relationships between foliar nitrogen (N) and phosphorus (P), leaf mass per area (LMA), and photosynthesis. However, how environmental conditions mediate LES trait interrelationships, particularly at large biospheric scales, remains unknown because of a lack of spatially explicit data, which ultimately limits our understanding of ecosystem processes, such as primary productivity and biogeochemical cycles. We used airborne imaging spectroscopy and geospatial modeling to generate, to our knowledge, the first biospheric maps of LES traits, here centered on 76 million ha of Andean and Amazonian forest, to assess climatic and geophysical determinants of LES traits and their interrelationships. Elevation and substrate were codominant drivers of leaf trait distributions. Multiple additional climatic and geophysical factors were secondary determinants of plant traits. Anticorrelations between N and LMA followed general LES theory, but topo-edaphic conditions strongly mediated and, at times, eliminated this classic relationship. We found no evidence for simple P-LMA or N-P trade-offs in forest canopies; rather, we mapped a continuum of N-P-LMA interactions that are sensitive to elevation and temperature. Our results reveal nested climatic and geophysical filtering of LES traits and their interrelationships, with important implications for predictions of forest productivity and acclimation to rapid climate change.

Species interactions are susceptible to anthropogenic changes in ecosystems, but this has been poorly investigated in a spatially explicit manner in the case of plant parasitism, such as the omnipresent hemiparasitic mistletoe-host plant interactions. Analyzing such interactions at a large spatial scale may advance our understanding of parasitism patterns over complex landscapes. Combining high-resolution airborne imaging spectroscopy and LiDAR, we studied hemiparasite incidence within and among tree host stands to examine the prevalence and spatial distribution of hemiparasite load in ecosystems. Specifically, we aimed to assess: (1) detection accuracy of mistletoes on their oak hosts; (2) hemiparasitism prevalence within host tree canopies depending on tree height, and (3) spatial variation in hemiparasitism across fragmented woodlands, in a low-diversity mediterranean oak woodland in California, USA. We identified mistletoe infestations with 55-96% accuracy, and detected significant differences in remote-sensed spectra between oak trees with and without mistletoe infestation. We also found that host canopy height had little influence on infestation degree, whereas landscape-level variation showed consistent; non-random patterns: isolated host trees had twice the infestation load than did trees located at the core of forest fragments. Overall, we found that canopy exposure (i.e., lower canopy density or proximity to forest edge) is more important than canopy height for mistletoe infestation, and that by changing landscape structure, parasitic prevalence increased with woodland fragmentation. We conclude that reducing fragmentation in oak woodlands will minimize anthropogenic impact on mistletoe infestation at the landscape level. We argue that advanced remote sensing technology can provide baselines to quantitatively analyze and monitor parasite-host trajectories in light of global environmental change, and that this is a promising approach to be further tested in other temperate and tropical forests.

We used aerial photography from 1954 and airborne LiDAR and imaging spectroscopy from 2008 to infer changes in extent and location of tallstature woody vegetation in 127 km2 of subalpine dry forest on the island of Hawai'i (Pōhakuloa Training Area), and to identify 25.8 km2 of intact woody vegetation for restoration and management. Total cover of woody vegetation was 54.7 km2 in 1954 and 58.6 km2 in 2008. Approximately 28.9 km2 underwent woody vegetation change (22.7%) between 1954 and 2008. Increases in woody vegetation cover occurred in 16.4 km2, and 12.5 km2 represented reduction of woody vegetation cover (12.9% and 9.8% of the 127 km2 study area, respectively). Our findings suggest that 3.9 km2 (3.0%) experienced a net increase in woody vegetation cover between 1954 and 2008. Spatial patterns suggest that fires may be the primary driver of reductions in woody vegetation cover. Increases could be due to regeneration of dry forest trees or measurement errors associated with historical imagery....

High-resolution mapping of tropical forest carbon stocks can assist forest management and improve implementation of large-scale carbon retention and enhancement programs. Previous high-resolution approaches have relied on field plot and/or light detection and ranging (LiDAR) samples of aboveground carbon density, which are typically upscaled to larger geographic areas using stratification maps. Such efforts often rely on detailed vegetation maps to stratify the region for sampling, but existing tropical forest maps are often too coarse and field plots too sparse for high-resolution carbon assessments. We developed a top-down approach for high-resolution carbon mapping in a 16.5 million ha region (> 40%) of the Colombian Amazon – a remote landscape seldom documented. We report on three advances for large-scale carbon mapping: (i) employing a universal approach to airborne LiDAR-calibration with limited field data; (ii) quantifying environmental controls over carbon densities; and (iii) developing stratification- and regression-based approaches for scaling up to regions outside of LiDAR coverage. We found that carbon stocks are predicted by a combination of satellite-derived elevation, fractional canopy cover and terrain ruggedness, allowing upscaling of the LiDAR samples to the full 16.5 million ha region. LiDAR-derived carbon maps have 14% uncertainty at 1 ha resolution, and the regional map based on stratification has 28% uncertainty in any given hectare. High-resolution approaches with quantifiable pixel-scale uncertainties will provide the most confidence for monitoring changes in tropical forest carbon stocks. Improved confidence will allow resource managers and decision makers to more rapidly and effectively implement actions that better conserve and utilize forests in tropical regions.

Remote sensing of canopy chemistry could greatly advance the study and monitoring of functional processes and biological diversity in humid tropical forests. Imaging spectroscopy has contributed to canopy chemical remote sensing, but efforts to develop general, globally-applicable approaches have been limited by sparse and inconsistent field and laboratory data, and lacking analytical methods. We analyzed leaf hemispherical reflectance and transmittance spectra, along with a 21-chemical portfolio, taken from 6136 fully sunlit humid tropical forest canopies, and developed an up-scaling method using a combination of canopy radiative transfer, chemometric and high-frequency noise modeling. By integrating these steps, we found that the accuracy and precision of multi-chemical remote sensing of tropical forest canopies varies by leaf constituent and wavelength range. Under conditions of varying canopy structure and spectral noise, photosynthetic pigments, water, nitrogen, cellulose, lignin, phenols and leaf mass per area (LMA) are accurately estimated using visible-to-shortwave infrared spectroscopy (VSWIR; 400–2500 nm). Phosphorus and base cations are retrieved with lower yet significant accuracy. We also find that leaf chemical properties are estimated far more consistently, and with much higher precision and accuracy, using the VSWIR range rather than the more common and limited visible to near-infrared range (400–1050 nm; VNIR). While VNIR spectroscopy proved accurate for predicting foliar LMA, photosynthetic pigments and water, VSWIR spectra provided accurate estimates for three times the number of canopy traits. These global results proved to be independent of site conditions, taxonomic composition and phylogenetic history, and thus they should be broadly applicable to multi-chemical mapping of humid tropical forest canopies. The approach developed and tested here paves the way for studies of canopy chemical properties in humid tropical forests using the next generation of airborne and space-based high-fidelity imaging spectrometers.

Current markets and international agreements for reducing emissions from deforestation and forest degradation (REDD) rely on carbon (C) monitoring techniques. Combining field measurements, airborne light detection and ranging (LiDAR)-based observations, and satellite-based imagery, we developed a 30-meter-resolution map of aboveground C density spanning 40 vegetation types found on the million-hectare Island of Hawaii. We estimate a total of 28.3 teragrams of C sequestered in aboveground woody vegetation on the island, which is 56% lower than Intergovernmental Panel on Climate Change estimates that do not resolve C variation at fine spatial scales. The approach reveals fundamental ecological controls over C storage, including climate, introduced species, and land-use change, and provides a fourfold decrease in regional costs of C measurement over field sampling alone.

The Brazilian savanna biome, known locally as the Cerrado, with an area of about 2 million km2 and marked by a conspicuous seasonality, comprises a vertically structured mosaic of ecosystem types, ranging from grassland to tropical dry forests. The Cerrado is a major agricultural frontier in Brazil, with nearly 50% of its original vegetative cover already converted to pastures and crop fields. Such large-scale conversion has severely affected regional runoff, river discharge and the atmosphere water transfer from soil reservoirs through vegetation. In this study, we used multitemporal Earth Observing-1 (EO-1) Hyperion hyperspectral imagery to derive canopy water content (validated by ground truth measurements), whose estimates were regionally extrapolated, over the entire Cerrado biome, based on the normalized difference vegetation index (NDVI) Moderate Resolution Imaging Spectroradiometer (MODIS) product MOD13Q1. MODIS-based canopy-level equivalent water thickness (EWTC) values were significantly distinc...

Millions of people rely on savannas for ecosystem services, such as the provision of grazing and fuel wood, so it is important to determine the extent to which utilization affects woody vegetation resources. Using airborne LiDAR from the Carnegie Airborne Observatory (CAO), we quantified and compared tree canopy cover and height distributions between areas of contrasting management in the Lowveld savanna region of South Africa – a region connecting communal landscapes with heavy utilization (especially fuel wood harvesting) to fully protected public (Kruger National Park – KNP) and private reserves (SabiSand Game Reserve – SSGR) that conserve biodiversity. Differences in total woody vegetation cover and cover within functional height classes (1–2 m, 2–3 m, 3–5 m, 5–7 m and >7 m) were investigated between 7 sites located within (i) conservation areas (in KNP, SSGR), (ii) communal rangelands or (iii) cultivated fields in communal areas. The impact of human utilization on wood resources in the communal areas varied widely between sites. Heavy utilization on gabbro substrate greatly reduced total woody cover of the rangelands, while two other communal rangelands that were presumably less intensively utilised had double the total woody cover of conservation areas. Rangelands and fields in most of the communal sites had more vegetation cover in the 5–7 m and >7 m classes than most of the conservation sites, presumably due to the absence of elephants in communal rangelands and the active preservation of large fruiting trees. On granite substrates, which account for the majority of the study area, there was a 50% reduction in woody cover below 5 m in communal rangelands. Although large trees were clearly being conserved in communal rangelands and fields, there was a relatively low cover of vegetation below 5 m, which raise doubts about recruitment and long-term sustainability of the tree resources. These results in conjunction with other studies based on the CAO LiDAR data for experimental burn plots and large mammal exclosures in KNP, suggest that communal land use on granite substrates have a higher impact on the woody cover below 5 m than both elephants and fire.

Termite mounds form islands of fertility in savanna landscapes and create foraging hotspots for herbivores, but the magnitude and spatial extent of these influences is unknown. We mapped terrain, termite mound and woody vegetation three-dimensional (3-D) structure at 56 cm resolution across a large-scale (254 ha), long-term (34 years) herbivore exclusion experiment in the Kruger National Park, with the Carnegie Airborne Observatory (CAO). We compared vegetation 3-D structure in areas protected from herbivores with those accessible to herbivores, both on termite mounds and in the landscape matrix between termite mounds. Termite mound density was 1.1 ha −1 across the study area and mound size did not differ between protected and accessible areas. Woody vegetation canopy cover was ∼100% greater on protected than accessible mounds, but was only ∼20% greater in the protected inter-mound matrix when compared to the accessible matrix. Woody canopy height class distributions differed significantly between protected and accessible areas, with the tallest vegetation (>10 m) occurring on protected termite mounds. The impacts of herbivore browsing were evident at distances of up to 20 m from termite mound centres. Spatial analysis of mound distribution revealed that the sphere of termite mound influence constitutes ∼20% of the total landscape. Termite influences on herbivore browsing operate at scales much larger than the spatial extent of their mound building activities.

Efforts to mitigate climate change through the Reduced Emissions from Deforestation and Degradation (REDD) depend on mapping and monitoring of tropical forest carbon stocks and emissions over large geographic areas. With a new integrated use of satellite imaging, airborne light detection and ranging, and field plots, we mapped aboveground carbon stocks and emissions at 0.1-ha resolution over 4.3 million ha of the Peruvian Amazon, an area twice that of all forests in Costa Rica, to reveal the determinants of forest carbon density and to demonstrate the feasibility of mapping carbon emissions for REDD. We discovered previously unknown variation in carbon storage at multiple scales based on geologic substrate and forest type. From 1999 to 2009, emissions from land use totaled 1.1% of the standing carbon throughout the region. Forest degradation, such as from selective logging, increased regional carbon emissions by 47% over deforestation alone, and secondary regrowth provided an 18% offset against total gross emissions. Very high-resolution monitoring reduces uncertainty in carbon emissions for REDD programs while uncovering fundamental environmental controls on forest carbon storage and their interactions with land-use change.

Uncertainties in burning efficiency (BE) estimates can lead to large errors in fire emission quantification (from 23% to 46%). One of the main causes of these errors is the spatial variability of fuel consumption within burned areas. This paper studies whether burn severity (BS) maps can be used to improve BE assessment. A burn severity map of two large fires in California was obtained by inverting a simulation model constrained by post-fire observations from Landsat TM imagery. Model output values of BS were validated against field measurements, obtaining a high correlation (R2 = 0.85) and low errors (Root Mean Square Error, RMSE = 0.14) throughout a wide range of BS levels. The BS map obtained was then used to adjust BE reference values per vegetation type found in the area before the fire. The adjusted burning efficiency (BEadj) was compared to the burned biomass, which was estimated by subtracting vegetation indices from pre- and post-fire images. Results showed a high correlation for conifers (R2 = 0.75) and hardwoods (R2 = 0.73), and a moderate correlation (R2 ∼ 0.5) for shrubs and grasslands. In general, for all vegetation types BEadj performed better (R2 = 0.4–0.75) than literature-based BE (R2

Rapid ‘Ōhi‘a Death (ROD) is a disease aggressively killing large numbers of Metrosideros polymorpha (‘ōhi‘a), a native keystone tree species on Hawaii Island. This loss threatens to deeply alter the biological make-up of this unique island ecosystem. Spatially explicit information about the present and past advancement of the disease is essential for its containment; yet, currently such data are severely lacking. To this end, we used the Carnegie Airborne Observatory to collect Laser-Guided Imaging Spectroscopy data and high-resolution digital imagery across >500,000 ha of Hawaii Island in June–July 2017. We then developed a method to map individual tree crowns matching the symptoms of both active (brown; desiccated ‘ōhi‘a crowns) and past (leafless tree crowns) ROD infection using an ensemble of two distinct machine learning approaches. Employing a very conservative classification scheme for minimizing false-positives, model sensitivity rates were 86.9 and 82.5, and precision rates were 97.4 and 95.3 for browning and leafless crowns, respectively. Across the island of Hawaii, we found 43,134 individual crowns suspected of exhibiting the active (browning) stage of ROD infection. Hotspots of potential ROD infection are apparent in the maps. The peninsula on the eastern side of Hawaii known as the Puna district, where the ROD outbreak likely originated, contained a particularly high density of brown crown detections. In comparison, leafless crown detections were much more numerous (547,666 detected leafless crowns in total) and more dispersed across the island. Mapped hotspots of likely ROD incidence across the island will enable scientists, administrators, and land managers to better understand both where and how ROD spreads and how to apply limited resources to limiting this spread.

The relative importance of fire and herbivory on vegetation structure has been the subject of much debate in savanna ecology. Fire regime and herbivore numbers are two key variables that managers of protected areas can manipulate to meet their conservation objectives. We deployed a new airborne remote sensing system (Carnegie Airborne Observatory) to the Kruger National Park (KNP), South Africa, to map a unique herbivore/fire exclusion experiment on basaltic soils. We collected high resolution (56 cm) three-dimensional (3-D) vegetation structural data over areas that have been protected from herbivores (34 yr) and/or fire (7 yr), as well as those exposed to both disturbance agents. Canopy height distribution, as well as the distribution of foliage within the vertical canopy profile, differed significantly between all treatments and between each treatment and the control area (Kolmogorov–Smirnov, p 9 m) and up to 66 times more small tree canopy (3–6 m). Fire restricted growth of vegetation in the 0–3 m height range, both in the absence and presence of herbivores. Our findings highlight the active role that conservation managers can play in modifying vegetation structure and heterogeneity through herbivore and fire management, as well as the value of 3-D remote sensing for the assessment of conservation management outcomes.

The negative effects of biological invasion are often the focus of ecological studies, but few have considered potential positive impacts, such as increased carbon storage, resulting from invasion. We combined airborne imaging spectrometer and LiDAR (light detection and ranging) observations with field measurements to assess if the highly invasive nitrogen-fixing tree Morella faya alters canopy 3-D structure and aboveground biomass (AGB) along a 1,500 mm precipitation gradient in Hawaii. Airborne analysis of canopy water content, leaf nitrogen concentration, fractional canopy cover, and vegetation height facilitated mapping of native- and Morella-dominated canopies in rainforest, woodland–savanna and shrubland ecosystems, with Morella detection errors ranging from 0 to 13.4%. Allometric equations were developed to relate the combined LiDAR and spectral data to field-based AGB estimates (r2 = 0.97, P < 0.01), and to produce a map of biomass stocks throughout native and invaded ecosystems. The structure of the invasive Morella canopies varied by ecosystem type, and the invader shaded out native understory plants in rainforest zones. Despite a 350% increase in AGB going from shrubland to rainforest, Morella did not increase average AGB in any ecosystem it invaded. Furthermore, spatial distributions of biomass indicated that Morella decreased maximum AGB in the woodland–savanna ecosystems. We conclude that Morella tree invasion does not enhance aboveground carbon stocks in any ecosystem it invades in Hawaii, thereby minimizing its contribution to this potentially important ecosystem service. We also found that the fusion of spectral and LiDAR remote sensing provided canopy chemical and structural data facilitating a landscape assessment of how biological invasion alters on carbon stocks and other ecosystem properties.

The environmental and biotic factors affecting spatial variation in canopy three-dimensional (3-D) structure and aboveground tree biomass (AGB) are poorly understood in tropical rain forests. We combined field measurements and airborne light detection and ranging (lidar) to quantify 3-D structure and AGB across a 5,016 ha rain forest reserve on the northeastern flank of Mauna Kea volcano, Hawaii Island. We compared AGB among native stands dominated by Metrosideros polymorpha found along a 600–1800 m elevation/climate gradient, and on three substrate-age classes of 5, 20, and 65 kyr. We also analyzed how alien tree invasion, canopy species dominance and topographic relief influence AGB levels. Canopy vertical profiles derived from lidar measurements were strong predictors (r 2 = 0.78) of AGB across sites and species. Mean AGB ranged from 48 to 363 Mg ha−1 in native forest stands. Increasing elevation corresponded to a 53–84% decrease in AGB levels, depending upon substrate age. Holding climate constant, changes in substrate age from 5 to 65 kyr corresponded to a 23–53% decline in biomass. Invasion by Psidium cattleianum and Ficus rubiginosa trees resulted in a 19–38% decrease in AGB, with these carbon losses mediated by substrate age. In contrast, the spread of former plantation tree species Fraxinus uhdei corresponded to a 7- to 10-fold increase in biomass. The effects of topographic relief at both local and regional scales were evident in the AGB maps, with poorly drained terrain harboring 76% lower biomass than forests on well-drained relief. Our results quantify the absolute and relative importance of environmental factors controlling spatial variation in tree biomass across a rain forest landscape, and highlight the rapid changes in carbon storage incurred following biological invasion.

Forest fragmentation results from deforestation and disturbance, with subsequent edge effects extending deep into remaining forest areas. No study has quantified the effects of both deforestation and selective logging, separately and combined, on forest fragmentation and edge effects over large regions. The main objectives of this study were to: (1) quantify the rates and extent of forest fragmentation from deforestation and logging within the Brazilian Amazon, and (2) contextualize the spatio-temporal dynamics of this forest fragmentation through a literature review of potential ecological repercussions of edge creation. Using GIS and remote sensing, we quantified forest fragmentation ‐ defined as both increases in the forest edge-to-area ratio and number of forest fragments ‐ and edge-effected forest occurring from these activities across more than 1.1 million km 2 of the Brazilian Amazon from 1999 to 2002. Annually, deforestation and logging generated 32,000 and 38,000 km of new forest edge while increasing the edge-to-area ratio of remaining forest by 0.14 and 0.15, respectively. Combined deforestation and logging increased the edge-to-area ratio of remaining forest by 65% over our study period, while generating 5539 and 3383 new forest fragments, respectively. Although we found that 90% of individual forest fragments were smaller than 4 km 2 , we also found that 50% of the remaining intact forests were located in contiguous forest areas greater than 35,000 km 2 . We then conducted a literature review documenting 146 edge effects and found that these penetrated to a median distance of 100 m, a distance encompassing 6.4% of all remaining forests in our study region in the year 2002, while 53% of forests were located within two km of an edge. Annually deforestation and logging increased the proportion of edge-forest by 0.8% and 3.1%, respectively. As a result of both activities, the total proportion of edge-forest increased by 2.6% per year, while the proportion within 100-m increased by 0.5%. Over our study period, deforestation resulted in an additional 3000 km 2 of edge-forest, whereas logging generated 20,000 km 2 , as it extended deep

Detection and mapping of invasive species is an important component of conservation and management efforts in Hawai'i, but the spectral separability of native, introduced, and invasive species has not been established. We used high spatial resolution airborne imaging spectroscopy to analyze the canopy hyperspectral reflectance properties of 37 distinct species or phenotypes, 7 common native and 24 introduced tree species, the latter group containing 12 highly invasive species. Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) reflectance and derivative-reflectance signatures of Hawaiian native trees were generically unique from those of introduced trees. Nitrogen-fixing trees were also spectrally unique from other groups of non-fixing trees. There were subtle but significant differences in the spectral properties of highly invasive tree species in comparison to introduced species that do not proliferate across Hawaiian ecosystems. The observed differences in canopy spectral signatures were linked to relative differences in measured leaf pigment (chlorophyll, carotenoids), nutrient (N, P), and structural (specific leaf area; SLA) properties, as well as to canopy leaf area index. These leaf and canopy properties contributed variably to the spectral separability of the trees, with wavelength-specific reflectance and absorption features that overlapped, but which were unique from one another. A combination of canopy reflectance from 1125–2500 nm associated with leaf and canopy water content, along with pigment-related absorption features (reflectance derivatives) in the 400–700 nm range, was best for delineating native, introduced, and invasive species. There was no single spectral region that always defined the separability of the species groups, and thus the full-range (400–2500 nm) spectrum was highly advantageous in differentiating these groups. These results provide a basis for more detailed studies of invasive species in Hawai'i, along with more explicit treatment of the biochemical properties of the canopies and their prediction using imaging spectroscopy.