Airborne Imaging Spectroscopy Research Articles

Spectra-phenology integration for high-resolution, accurate, and scalable mapping of foliar functional traits using time-series Sentinel-2 data

Foliar functional traits are essential for understanding plant adaptation strategies and ecosystem function. Due to limited in-situ observational data, there is a growing interest in upscaling these traits from field sites to regional and global levels. However, limitations persist: (1) global/national scale upscaling that relies on plant functional type (PFT) maps, environmental variables or coarse resolution multispectral images, which fail to capture local-scale trait variability; (2) airborne imaging spectroscopy that enables high-resolution and accurate mapping but is restricted to site scale and is costly; and (3) multispectral satellites like Sentinel-2 that offer global coverage but have limited spectral bands and resolution. While previous research has demonstrated the connection between traits and vegetation phenology, our study seeks to build upon this foundation by further exploring the integration of phenological information for large-scale trait prediction. We examined the integration of Sentinel-2 data with its time series (for phenology information) to map 12 foliar functional traits across 14 National Ecological Observatory Network (NEON) sites in the eastern United States. Our results show that time-series Sentinel-2 models effectively capture the variance in these 12 traits (R2 = 0.60–0.80) when compared with benchmark trait data generated by state-of-the-art airborne imaging spectroscopy. The models adequately capture considerable trait variations observed within sites and PFTs. Our approach outperforms existing methods that rely on environmental variables, or a single Sentinel-2 image as predictors across examined NEON sites in eastern United States. Interestingly, including environmental variables in our models does not significantly improve predictive power. Further analysis reveals that a ‘fast-slow’ principal axis predominantly explains the covariation in Enhanced Vegetation Index amplitude (a proxy for leaf longevity), leaf mass per area, and leaf nitrogen content across PFTs. This finding highlights the importance of incorporating phenological information for trait mapping and suggests a potential mechanism underlying these spectra-based models. Our proposed method, which simultaneously achieves high accuracy, large-scale scalability, and high spatial resolution, represents a promising avenue for future global trait mapping. Validation on a larger scale to fully realize its potential in addressing fundamental ecological questions will be a key future focus.

Variability in Symbiont Chlorophyll of Hawaiian Corals from Field and Airborne Spectroscopy

Corals are habitat-forming organisms on tropical and sub-tropical reefs, often displaying diverse phenotypic behaviors that challenge field-based monitoring and assessment efforts. Symbiont chlorophyll (Chl) is a long-recognized indicator of intra- and inter-specific variation in coral’s response to environmental variability and stress, but the quantitative Chl assessment of corals at the reef scale continues to prove challenging. We integrated field, airborne, and laboratory techniques to test and apply the use of reflectance spectroscopy for in situ and reef-scale estimation of Chl a and Chl c2 concentrations in a shallow reef environment of Kāne‘ohe Bay, O‘ahu. High-fidelity spectral signatures (420–660 nm) derived from field and airborne spectroscopy quantified Chl a and Chl c2 concentrations with demonstrable precision and accuracy. Airborne imaging spectroscopy revealed a 10-fold range of Chl concentrations across the reef ecosystem. We discovered a differential pattern of Chl a and Chl c2 use in symbiont algae in coexisting corals indicative of a physiological response to decreasing light levels with increasing water depth. The depth-dependent ratio of Chl c2:a indicated the presence of two distinct light-driven habitats spanning just 5 m of water depth range. Our findings provide a pathway for further study of coral pigment responses to environmental conditions using field and high-resolution airborne imaging spectroscopy.

A general methodology for the quantification of crop canopy nitrogen across diverse species using airborne imaging spectroscopy

Accurate monitoring of crop nitrogen (N) across spatial and temporal scales is a fundamental goal for meeting precision agriculture requirements and promoting sustainable agriculture. The planning and implementation of several spaceborne imaging spectroscopy missions in recent years holds great promise for such large scale and intricate monitoring. Several N retrieval models have been developed for specific crop species, but a generalized model across diverse species is lacking. By leveraging imaging spectroscopy data collected by the Global Airborne Observatory (GAO), and leaf samples collected from commercial and research farms, we used partial least squares regression to calibrate and validate the retrieval of mass-based crop canopy N concentrations across diverse species and several major agricultural regions in the contiguous United States. The performance statistics indicated high precision and accuracy of the model results, suggesting that the development of a generalized N retrieval model is possible (R2: 0.78; RMSE: 0.49% N). Maps derived from GAO data provided quantitative crop N information at fine spatial resolution (i.e., 0.6 m), capturing both inter- and intra-species variations across agricultural locations. The algorithm was also successfully tested on simulated moderate resolution (i.e., 30 m) imagery, corresponding to data to be collected by forthcoming spaceborne imaging spectroscopy missions. Imaging spectroscopy offers an effective approach to quantify crop N concentration that could be incorporated to promote sustainable agriculture and improve global food security.

Tundra fire increases the likelihood of methane hotspot formation in the Yukon–Kuskokwim Delta, Alaska, USA

Rapid warming in Arctic tundra may lead to drier soils in summer and greater lightning ignition rates, likely culminating in enhanced wildfire risk. Increased wildfire frequency and intensity leads to greater conversion of permafrost carbon to greenhouse gas emissions. Here, we quantify the effect of recent tundra fires on the creation of methane (CH4) emission hotspots, a fingerprint of the permafrost carbon feedback. We utilized high-resolution (∼25 m2 pixels) and broad coverage (1780 km2) airborne imaging spectroscopy and maps of historical wildfire-burned areas to determine whether CH4 hotspots were more likely in areas burned within the last 50 years in the Yukon–Kuskokwim Delta, Alaska, USA. Our observations provide a unique observational constraint on CH4 dynamics, allowing us to map CH4 hotspots in relation to individual burn events, burn scar perimeters, and proximity to water. We find that CH4 hotspots are roughly 29% more likely on average in tundra that burned within the last 50 years compared to unburned areas and that this effect is nearly tripled along burn scar perimeters that are delineated by surface water features. Our results indicate that the changes following tundra fire favor the complex environmental conditions needed to generate CH4 emission hotspots. We conclude that enhanced CH4 emissions following tundra fire represent a positive feedback that will accelerate climate warming, tundra fire occurrence, and future permafrost carbon loss to the atmosphere.

Classifying a Highly Polymorphic Tree Species across Landscapes Using Airborne Imaging Spectroscopy

Vegetation classifications on large geographic scales are necessary to inform conservation decisions and monitor keystone, invasive, and endangered species. These classifications are often effectively achieved by applying models to imaging spectroscopy, a type of remote sensing data, but such undertakings are often limited in spatial extent. Here we provide accurate, high-resolution spatial data on the keystone species Metrosideros polymorpha, a highly polymorphic tree species distributed across bioclimatic zones and environmental gradients on Hawai’i Island using airborne imaging spectroscopy and LiDAR. We compare two tree species classification techniques, the support vector machine (SVM) and spectral mixture analysis (SMA), to assess their ability to map M. polymorpha over 28,000 square kilometers where differences in topography, background vegetation, sun angle relative to the aircraft, and day of data collection, among others, challenge accurate classification. To capture spatial variability in model performance, we applied Gaussian process classification (GPC) to estimate the spatial probability density of M. polymorpha occurrence using only training sample locations. We found that while SVM and SMA models exhibit similar raw score accuracy over the test set (96.0% and 93.4%, respectively), SVM better reproduces the spatial distribution of M. polymorpha than SMA. We developed a final 2 m × 2 m M. polymorpha presence dataset and a 30 m × 30 m M. polymorpha density dataset using SVM classifications that have been made publicly available for use in conservation applications. Accurate, large-scale species classifications are achievable, but metrics for model performance assessments must account for spatial variation of model accuracy.

Scalable early detection of grapevine virus infection with airborne imaging spectroscopy.

The US wine and grape industry suffers $3B in damages and losses annually due to viral diseases such as Grapevine Leafroll-associated Virus Complex 3 (GLRaV-3). Current detection methods are labor intensive and expensive. GLRaV-3 undergoes a latent period in which the vines are infected but do not yet display visible symptoms, making it an ideal model to evaluate the scalability of imaging spectroscopy-based disease detection. We deployed the NASA Airborne Visible and Infrared Imaging Spectrometer Next Generation (AVIRIS-NG) to detect GLRaV-3 in Cabernet Sauvignon grapevines in Lodi, CA in September 2020. Foliage was removed from the vines as part of mechanical harvest soon after imagery acquisition. In both Sept. 2020 and 2021, industry collaborators scouted 317ac on a vine-by-vine basis for visible viral symptoms and collected a subset for molecular confirmation testing. Grapevines identified as visibly diseased in 2021, but not 2020, were assumed to have been latently infected at time of acquisition. We trained spectral models with random forest and synthetic minority oversampling technique to distinguish non-infected and GLRaV-3-infected grapevines. Non-infected and GLRaV-3 infected vines could be differentiated both pre- and post-symptomatically at 1m through 5m resolution. The best-performing models had 87% accuracy distinguishing between non-infected and asymptomatic vines, and 85% accuracy distinguishing between non-infected and asymptomatic + symptomatic vines. The importance of non-visible wavelengths suggests this capacity is driven by disease-induced changes to overall plant physiology. Our work sets a foundation for using the forthcoming hyperspectral satellite Surface Biology and Geology for regional disease monitoring.

Uncovering the hidden: Leveraging sub-pixel spectral diversity to estimate plant diversity from space

Remotely sensed spectral diversity has emerged as a promising proxy for plant diversity. However, spectral diversity approaches relate image spectra to plant community diversity by only incorporating variation among adjacent pixels, considering each pixel as a homogeneous entity composed of one class, and disregarding the within-pixel variability. Although such approaches might work for remotely-sensed data with fine spatial resolution, they might not be viable solutions to estimate plant diversity using coarse-resolution data from forthcoming spaceborne imagers. To address the limitations associated with spectral diversity approaches, we proposed a novel approach, known as endmember diversity, for remote estimation of plant diversity through quantifying spectral diversity at the sub-pixel level and taking into account the within-pixel variability. The approach consisted of deriving the number and abundance of distinct spectral entities within each pixel via spectral unmixing. In doing so, we considered the spectral signature of each pixel as a mixture of distinct spectral entities, commonly known as endmembers. We then used the per-pixel endmembers and their abundance to calculate different spectral diversity metrics for every pixel. We assessed the performance of the endmember diversity approach at estimating plant taxonomic and phylogenetic diversity based on two experiments using a simulated spectral dataset and a real-world spaceborne DESIS (DLR Earth Sensing Imaging Spectrometer) dataset. In both experiments, we found significant associations between endmember diversity and in situ plant diversity. Additionally, our method applied to DESIS data outperformed a conventional spectral diversity metric based on the coefficient of variation when applied to 1-m airborne imaging spectroscopy data. Collectively, our results demonstrate the capability of forthcoming spaceborne imagers to monitor local plant diversity.

Do beaver ponds increase methane emissions along Arctic tundra streams?

Beaver engineering in the Arctic tundra induces hydrologic and geomorphic changes that are favorable to methane (CH4) production. Beaver-mediated methane emissions are driven by inundation of existing vegetation, conversion from lotic to lentic systems, accumulation of organic rich sediments, elevated water tables, anaerobic conditions, and thawing permafrost. Ground-based measurements of CH4 emissions from beaver ponds in permafrost landscapes are scarce, but hyperspectral remote sensing data (AVIRIS-NG) permit mapping of ‘hotspots’ thought to represent locations of high CH4 emission. We surveyed a 429.5 km2 area in Northwestern Alaska using hyperspectral airborne imaging spectroscopy at ∼5 m pixel resolution (14.7 million observations) to examine spatial relationships between CH4 hotspots and 118 beaver ponds. AVIRIS-NG CH4 hotspots covered 0.539% (2.3 km2) of the study area, and were concentrated within 30 m of waterbodies. Comparing beaver ponds to all non-beaver waterbodies (including waterbodies >450 m from beaver-affected water), we found significantly greater CH4 hotspot occurrences around beaver ponds, extending to a distance of 60 m. We found a 51% greater CH4 hotspot occurrence ratio around beaver ponds relative to nearby non-beaver waterbodies. Dammed lake outlets showed no significant differences in CH4 hotspot ratios compared to non-beaver lakes, likely due to little change in inundation extent. The enhancement in AVIRIS-NG CH4 hotspots adjacent to beaver ponds is an example of a new disturbance regime, wrought by an ecosystem engineer, accelerating the effects of climate change in the Arctic. As beavers continue to expand into the Arctic and reshape lowland ecosystems, we expect continued wetland creation, permafrost thaw and alteration of the Arctic carbon cycle, as well as myriad physical and biological changes.

Quantifying the Variation in Reflectance Spectra of Metrosideros polymorpha Canopies across Environmental Gradients

Imaging spectroscopy is a burgeoning tool for understanding ecosystem functioning on large spatial scales, yet the application of this technology to assess intra-specific trait variation across environmental gradients has been poorly tested. Selection of specific genotypes via environmental filtering plays an important role in driving trait variation and thus functional diversity across space and time, but the relative contributions of intra-specific trait variation and species turnover are still unclear. To address this issue, we quantified the variation in reflectance spectra within and between six uniform stands of Metrosideros polymorpha across elevation and soil substrate age gradients on Hawai‘i Island. Airborne imaging spectroscopy and light detection and ranging (LiDAR) data were merged to capture and isolate sunlit portions of canopies at the six M. polymorpha-dominated sites. Both intra-site and inter-site spectral variations were quantified using several analyses. A support vector machine (SVM) model revealed that each site was spectrally distinct, while Euclidean distances between site centroids in principal components (PC) space indicated that elevation and soil substrate age drive the separation of canopy spectra between sites. Coefficients of variation among spectra, as well as the intrinsic spectral dimensionality of the data, demonstrated the hierarchical effect of soil substrate age, followed by elevation, in determining intra-site variation. Assessments based on leaf trait data estimated from canopy reflectance resulted in similar patterns of separation among sites in the PC space and distinction among sites in the SVM model. Using a highly polymorphic species, we demonstrated that canopy reflectance follows known ecological principles of community turnover and thus how spectral remote sensing addresses forest community assembly on large spatial scales.

Evaluating individual tree species classification performance across diverse environments

Vegetation species mapping using airborne imaging spectroscopy yields accurate results and is important for advancing conservation objectives and biogeographic studies. As these data become more readily available owing to the upcoming launch of spaceborne imaging spectrometers, it is necessary to understand how these data can be used to consistently classify species across large geographic scales. However, few studies have attempted to map species across multiple ecosystems; therefore, little is known regarding the effect of intra-specific variation on the mapping of a single species across a wide range of environments and among varying backgrounds of other non-target species. To explore this effect, we developed and tested species classification models for Metrosideros polymorpha, a highly polymorphic canopy species endemic to Hawai’i, which is found in a diverse array of ecosystems. We compared the accuracies of support vector machine (SVM) and random forest models trained on canopy reflectance data from each of eight distinct ecosystems (ecosystem-specific) and a universal model trained on data from all ecosystems. When applied to ecosystem-specific test datasets, the ecosystem-specific models outperformed the universal model; however, the universal model retained high (>81%) accuracies across all ecosystems. Additionally, we found that models from ecosystems with broad variation in M. polymorpha canopy traits, as estimated using chemometric equations applied to canopy spectra, accurately predicted M. polymorpha in other ecosystems. While species classifications across ecosystems can yield accurate results, these results will require sampling procedures that capture the intra-specific variation of the target species.

Integrating very-high-resolution UAS data and airborne imaging spectroscopy to map the fractional composition of Arctic plant functional types in Western Alaska

Widespread changes in vegetation cover and composition are driving strong impacts on Arctic ecosystem functioning and global climate feedbacks. An accurate characterization of tundra vegetation composition is required to understand how the Arctic will respond to future climate change. However, quantifying tundra vegetation composition over large areas is challenging as commonly-used satellite observations are too coarse, spatially and spectrally, to differentiate low-lying tundra vegetation types. Recent airborne and spaceborne imaging spectroscopy platforms provide better data to characterize vegetation composition. Yet, our ability to characterize vegetation composition with imaging spectroscopy remains largely unexplored in the Arctic, particularly due to a lack of ground observations needed to train and test classification models. To address this problem, we collected very-high-resolution (VHR, ∼5 cm) unoccupied aerial system (UAS) imagery at three low-Arctic tundra sites located on the Seward Peninsula, western Alaska. In this paper, we examine the feasibility of integrating imagery from the UAS and the hyperspectral Airborne Visible/Infrared Imaging Spectrometer, Next Generation (AVIRIS-NG) airborne instrument to map the fractional composition of 12 key Arctic plant functional types (PFTs). To this end, we first mapped the 12 PFTs from our VHR UAS imagery using random forest classification. We then used these UAS-derived PFT maps as ground truth to develop partial least squares regression (PLSR) models to predict the fractional cover (FCover) of each PFT from AVIRIS-NG imagery. We evaluated the performance of our PLSR models using reserved UAS samples, as well as by mapping PFT FCover and dominant PFT for large tundra landscapes. Our results show that 1) Arctic PFTs can be effectively mapped using VHR UAS imagery, with overall accuracy between 86% and 92%, 2) when the UAS mapped PFTs were used to inform PLSR scaling models, the FCover of the 12 PFTs could be effectively estimated from AVIRIS-NG imagery with a mean absolute error (MAE) <0.13, and 3) our PLSR models outperformed traditional, fully constrained least-squares (FCLS) linear mixture analysis and produced high-quality, spatially contiguous PFT FCover and PFT maps that captured vegetation spatial patterns with similar accuracy to those developed from UAS imagery. The developed PLSR models have the potential to be broadly applied for quantifying vegetation composition with AVIRIS-NG images to help monitor tundra vegetation dynamics and improve process-based modeling of tundra ecosystems.

Mapping of understorey invasive plant species clusters of Lantana camara and Chromolaena odorata using airborne hyperspectral remote sensing

The present study highlights delineation of the distribution of understorey invasive plant species (IPS) in Mudumalai Tiger Reserve (MTR), Western Ghats, India using very high resolution airborne imaging spectroscopy images by evaluating the performance of a Multiple Endmember Spectral Mixture Analysis (MESMA). Lantana camara L. and Chromolaena odorata (L.) R.M.King & H.Rob. have imposed a serious threat to natural ecosystems in MTR by altering the ecosystem processes. The Airborne Visible-Infrared Imaging Spectrometer-Next Generation (AVIRIS-NG) based analysis delineated the distribution of understorey IPS, viz., L. camara and C. odorata in MTR, with an overall accuracy of 87 % (k = 0.84) and 84 % (k = 0.8), respectively, with minimal root-mean-square error (RMSE) of MESMA classification (0.001 to 0.0025). In MTR, both the IPS were distributed in large parts and in association with all dominant canopy tree species barring Eucalyptus species. L. camara was majorly found in the southern parts of central MTR with ‘high’ to ‘very high’ presence covering 23.09 % and 11.51 % of the total area of central MTR, respectively, in association with Terminalia anogeissiana Gere & Boatwr. and Tectona grandis L.f. Whereas, C. odorata was dominantly found with high (18.11 % area of central MTR) to very high (11.85 % area of central MTR) intensity in the south and southeastern parts within the vicinity of T. anogeissiana and Terminalia elliptica Willd. The study highlights the utilization of high spectral agility in mapping of understorey IPS through spectral differentiation of the patchy distribution of IPS and its intermixing with other local species may be helpful in ecological restoration, conservation of biodiversity, and provisioning long-term management of ecosystem services in highly invaded tropical forests.

Genetic constraints on temporal variation of airborne reflectance spectra and their uncertainties over a temperate forest

Remote sensing enhances large-scale biodiversity monitoring by overcoming temporal and spatial limitations of ground-based measurements and allows assessment of multiple plant traits simultaneously. The total set of traits and their variation over time is specific for each individual and can reveal information about the genetic composition of forest communities. Measuring trait variation among individuals of one species continuously across space and time is a key component in monitoring genetic diversity but difficult to achieve with ground-based methods. Remote sensing approaches using imaging spectroscopy can provide high spectral, spatial, and temporal coverage to advance the monitoring of genetic diversity, if sufficient relation between spectral and genetic information can be established.We assessed reflectance spectra from individual Fagus sylvatica L. (European beech) trees acquired across eleven years from 69 flights of the Airborne Prism Experiment (APEX) above the same temperate forest in Switzerland. We derived reflectance spectra of 68 canopy trees and correlated differences in these spectra with genetic differences derived from microsatellite markers among the 68 individuals. We calculated these correlations for different points in time, wavelength regions and relative differences between wavelength regions. High correlations indicate high spectral-genetic similarities. We then tested the influence of environmental variables obtained at temporal scales from days to years on spectral-genetic similarities. We performed an uncertainty propagation of radiance measurements to provide a quality indicator for these correlations.We observed that genetically similar individuals had more similar reflectance spectra, but this varied between wavelength regions and across environmental variables. The short-wave infrared regions of the spectrum, influenced by water absorption, seemed to provide information on the population genetic structure at high temperatures, whereas the visible part of the spectrum, and the near-infrared region affected by scattering properties of tree canopies, showed more consistent patterns with genetic structure across longer time scales. Correlations of genetic similarity with reflectance spectra similarity were easier to detect when investigating relative differences between spectral bands (maximum correlation: 0.40) than reflectance data (maximum correlation: 0.33). Incorporating uncertainties of spectral measurements yielded improvements of spectral-genetic similarities of 36% and 20% for analyses based on single spectral bands, and relative differences between spectral bands, respectively.This study highlights the potential of dense multi-temporal airborne imaging spectroscopy data to detect the genetic structure of forest communities. We suggest that the observed temporal trajectories of reflectance spectra indicate physiological and possibly genetic constraints on plant responses to environmental change.

Sensitivity of modeled snow grain size retrievals to solar geometry, snow particle asphericity, and snowpack impurities

Abstract. Snow grain size is an important metric to determine snow age and metamorphism, but it is difficult to measure. The effective grain size can be derived from spaceborne and airborne radiance measurements due to strong attenuation of near-infrared energy by ice. Consequently, a snow grain size inversion technique that uses hyperspectral radiances and exploits variations in the 1.03 µm ice absorption feature was previously developed for use with airborne imaging spectroscopy. Previous studies have since demonstrated the effectiveness of the technique, though there has yet to be a quantitative assessment of the retrieval sensitivity to snowpack impurities, ice particle shape, or solar geometry. In this study, we use the Snow, Ice, and Aerosol Radiative (SNICAR) model and a Monte Carlo photon tracking model to examine the sensitivity of snow grain size retrievals to changes in dust and black carbon content, anisotropic reflectance, changes in solar illumination angle (θ0), and scattering asymmetry parameter (g) associated with different particle shapes. Our results show that changes in these variables can produce large grain size errors, especially when the effective grain size exceeds 500 µm. Dust content of 1000 ppm induces errors exceeding 800 µm, with the highest biases associated with small particles. Aspherical ice particles and perturbed solar zenith angles produce maximum biases of ∼540 µm and ∼400 µm, respectively, when spherical snow grains and θ0=60∘ are assumed in the generation of the retrieval calibration curve. Retrievals become highly sensitive to viewing angle when reflectance is anisotropic, with biases exceeding 1000 µm in extreme cases. Overall, we show that a more detailed understanding of snowpack state and solar geometry improves the precision when determining snow grain size through hyperspectral remote sensing.

A non-destructive estimation of chlorophyll-a and -b over different crops using airborne imaging spectroscopy observations

Non-destructive estimation of chlorophyll-a (Chl-a) and chlorophyll-b (Chl-b), is required for assessment and understanding of vegetation structural and functional dynamics. We have developed an empirical model for estimating Chl-a and Chl-b for different crops using Airborne Visible-Infrared Imaging Spectrometer Next Generation (AVIRIS-NG) data. Second derivative reflectance at 672 nm and 587 nm has shown maximum sensitivity towards Chl-a and Chl-b content, respectively. Separate linear models within a bootstrapped resampling framework were developed. During calibration, mean R2 of 0.46 and 0.51 were obtained for Chl-a and -b respectively. While during model validation, in both the cases mean R2 was found to be 0.3. This work shows the capability of airborne hyperspectral imaging data in segregating chlorophyll types using a simplistic modeling framework.

Genus-Level Mapping of Invasive Floating Aquatic Vegetation Using Sentinel-2 Satellite Remote Sensing

Invasive floating aquatic vegetation negatively impacts wetland ecosystems and mapping this vegetation through space and time can aid in designing and assessing effective control strategies. Current remote sensing methods for mapping floating aquatic vegetation at the genus level relies on airborne imaging spectroscopy, resulting in temporal gaps because routine hyperspectral satellite coverage is not yet available. Here we achieved genus level and species level discrimination between two invasive aquatic vegetation species using Sentinel 2 multispectral satellite data and machine-learning classifiers in summer and fall. The species of concern were water hyacinth (Eichornia crassipes) and water primrose (Ludwigia spp.). Our classifiers also identified submerged and emergent aquatic vegetation at the community level. Random forest models using Sentinel-2 data achieved an average overall accuracy of 90%, and class accuracies of 79–91% and 85–95% for water hyacinth and water primrose, respectively. To our knowledge, this is the first study that has mapped water primrose to the genus level using satellite remote sensing. Sentinel-2 derived maps compared well to those derived from airborne imaging spectroscopy and we also identified misclassifications that can be attributed to the coarser Sentinel-2 spectral and spatial resolutions. Our results demonstrate that the intra-annual temporal gaps between airborne imaging spectroscopy observations can be supplemented with Sentinel-2 satellite data and thus, rapidly growing/expanding vegetation can be tracked in real time. Such improvements have potential management benefits by improving the understanding of the phenology, spread, competitive advantages, and vulnerabilities of these aquatic plants.

Estimation of Canopy Structure of Field Crops Using Sentinel-2 Bands with Vegetation Indices and Machine Learning Algorithms

Leaf angle distribution (LAD), or the leaf mean tilt angle (MTA) capturing its central value, is used to quantify the direction of the leaf surface in a canopy and is one of the most important canopy structuraltraits. Combined with the other important structure parameter, leaf area index (LAI), LAD determines the light interception of a crop canopy. However, unlike LAI, only few studies have addressed the direct retrieval of LAD or MTA from remote sensing data. Recently, it has been shown that the red edge is a key spectral region where the effect of leaf angle on crop spectral reflectance can be separated from that of other structural variables. The Multispectral imager (MSI) onboard the Sentinel-2 (S2) satellite has two specially designed red-edge channels in this spectral region and thus can potentially be used for large-scale mapping of MTA at high spatial and temporal resolutions. Unfortunately, no field data on leaf angles at the scale of S2 pixel are available. Therefore, we simulated 5000 observations of different crops using the PROSAIL canopy reflectance model. Further, we used the MTA and LAI data of six crop species growing in 162 experimental plots in Finland and simulated their reflectance signal in S2 bands by resampling AISA airborne imaging spectroscopy data. Four common machine learning regression algorithms (random forest, support vector machine, multilayer perceptron network and partial least squares regression) were examined for retrieving canopy structure parameters, including leaf angle, from the simulated reflectances. Further, we analyzed the utility of 12 vegetation indices (VIs) well known to be sensitive to canopy structure for canopy structure estimation. Six of the studied indices used information from the visible part of the spectrum and the near infrared (NIR) while another six were selected to also utilize the red edge bands specific to S2. We found that S2 band 6 in the red edge had a strong correlation with MTA (R2 = 0.79 in model simulation and R2 = 0.87 in field measurements) but a low correlation with LAI (R2 = 0.07 in model simulation and R2= 0.06 in field measurements). Of the six red edge-based VIs, four (NDVIRE, CIRE, WDRVIRE and MSRRE) depended less on MTA than the visible NIR-based VIs and thus could be useful for estimating LAI for any LAD. The other two red edge-based VIs, IRECI and S2REP, had stronger correlations with MTA (R2 = 0.67 and 0.52, respectively) than LAI (R2 = 0.24 and 0.19, respectively). Additionally, MTA was accurately estimated (RMSE = 1.1–2.4° in model simulations and RMSE = 2.2–3.9° in field measurements) using the four 10 m spatial resolution bands with the RF, SVM and MLP algorithms, without information in the red edge. These promising results indicate the capability of S2 in accurately mapping the MTA of field crops on a large scale.

Mapped coral mortality and refugia in an archipelago-scale marine heat wave

Corals are a major habitat-building life-form on tropical reefs that support a quarter of all species in the ocean and provide ecosystem services to millions of people. Marine heat waves continue to threaten and shape reef ecosystems by killing individual coral colonies and reducing their diversity. However, marine heat waves are spatially and temporally heterogeneous, and so too are the environmental and biological factors mediating coral resilience during and following thermal events. This combination results in highly variable outcomes at both the coral bleaching and mortality stages of every event. This, in turn, impedes the assessment of changing reef-scale patterns of thermal tolerance or places of resistance known as reef refugia. We developed a large-scale, high-resolution coral mortality monitoring capability based on airborne imaging spectroscopy and applied it to a major marine heat wave in the Hawaiian Islands. While water depth and thermal stress strongly mediated coral mortality, relative coral loss was also inversely correlated with preheat-wave coral cover, suggesting the existence of coral refugia. Subsequent mapping analyses indicated that potential reef refugia underwent up to 40% lower coral mortality compared with neighboring reefs, despite similar thermal stress. A combination of human and environmental factors, particularly coastal development and sedimentation levels, differentiated resilient reefs from other more vulnerable reefs. Our findings highlight the role that coral mortality mapping, rather than bleaching monitoring, can play for targeted conservation that protects more surviving corals in our changing climate.

Assessing biodiversity from space: Impact of spatial and spectral resolution on trait-based functional diversity

Observing functional diversity continuously in time and space using satellite imagery forms the basis for studying impact, interactions, and feedback of environmental change mechanisms on ecosystems and biodiversity globally. Functional diversity of plant traits links ecosystem functioning and biodiversity. This work presents an approach to map and quantify functional diversity of physiological forest traits derived from 20 m Sentinel-2 data in a temperate forest ecosystem. We used two complementary data sources, namely high-resolution, as well as spatially resampled airborne imaging spectroscopy data and Sentinel-2 data, to ensure our methods support consistently mapping functional diversity from space. We retrieved three physiological traits related to forest health, stress, and potential productivity, namely chlorophyll, carotenoid, and water content, from airborne imaging spectroscopy and Sentinel-2 data using corresponding spectral indices as proxies. We analyzed changes in two functional diversity metrics, namely functional richness and divergence, at different spatial resolutions. Both functional diversity metrics depend on the size and number of pixels to derive functional diversity as a function of distance, leading to different interpretations. When mapping functional diversity using Sentinel-2 data, small-scale patterns <1.1 ha were no longer visible, implying a minimum calculation area with 60 m radius recommended for retrieval of functional diversity metrics. The spectrally convolved and spatially resampled airborne spectroscopy data and the native Sentinel-2 data were correlated with r = 0.747 for functional richness and r = 0.709 for divergence in a 3.1 ha neighborhood. Functional richness was more affected by the differences in trait maps between the acquisitions resulting from effects in illumination and topography compared with functional divergence. Further differences could be explained by varying illumination/observation effects and phenological status of the vegetation at acquisition. Our approach demonstrates the importance of spatial and spectral resolution when scaling diversity assessments from regional to continental scales.

Characterizing seasonal variation in foliar biochemistry with airborne imaging spectroscopy

Foliar biochemical traits are important indicators of ecosystem functioning and health that are impractical to characterize at large spatial and temporal scales using traditional measurements. However, comprehensive inventories of foliar traits are important for understanding ecosystem responses to anthropogenic and natural disturbances, as inputs into ecosystem process models, and for quantifying spatial variation in functional diversity. Imaging spectroscopy has been demonstrated as a valuable tool for developing maps of ecologically important foliar traits at large scales, but its application to mapping foliar traits over the course of the growing season has been limited. We collected high-resolution imaging spectroscopy data over Blackhawk Island, Wisconsin, USA at eight time points during the 2018 growing season (May – October). Using partial least squares regression (PLSR) we developed predictive models applicable to all dates to produce canopy-level maps of eight traits related to ecophysiological function: chlorophyll content, leaf mass per area and concentrations of calcium, nitrogen, phosphorus, potassium, phenolics and lignin. The accuracy of our models varied across traits (R2: 0.25–0.86); traits with well-defined absorption features were retrieved with high accuracy including chlorophyll (R2: 0.86; %RMSE: 11.0) and total phenolics (R2: 0.86; %RMSE: 11.0). We also assessed how well our models estimated biochemistry on novel species and new dates using a cross-validation analysis. Chlorophyll and total phenolics were well estimated across withheld dates and species, whereas calcium was estimated poorly on both withheld species (R2: 0.08) and dates (R2: 0.07). Our canopy-level maps of macronutrients (N, P and K) showed general trends of decreasing concentration over the course of the year, reflecting dilution by carbon-rich compounds during the growing season and resorption during senescence. Conversely, recalcitrant compounds including lignin and calcium increased until late summer, after which they stabilized. These results demonstrate the potential of current and proposed spaceborne imaging spectroscopy missions for mapping seasonal patterns in foliar biochemistry at a global scale.