Grassland foliar functional traits are closely linked to ecosystem functioning, biodiversity, and plant responses to environmental change. Hyperspectral remote sensing provides an efficient and non-destructive approach for mapping foliar traits, yet direct comparisons between UAV-based and airborne imaging spectroscopy remain limited. In this study, we evaluated the performance of UAV-based Nano and airborne Hyspex hyperspectral imagery for predicting ten foliar functional traits across experimental grassland plots at the Cedar Creek Ecosystem Science Reserve, USA. We further assessed the contributions of visible-to-near-infrared (VNIR) and shortwave infrared (SWIR) spectral regions, as well as the effects of spectral preprocessing approaches for minimizing confounding effects from canopy structure, illumination/viewing geometry, and soil background. Random Forest regression models were developed using plot-level average spectra derived from Nano and Hyspex imagery. Both UAV- and airborne-based imaging spectroscopy achieved moderate to high prediction accuracies for most foliar traits. High accuracies were obtained for non-structural carbohydrates (NSC), carotenoids, β-carotene, hemicellulose, and cellulose (R2 = 0.66–0.82; NRMSE = 6–10%), while moderate accuracies were achieved for nitrogen, chlorophyll, and xanthophylls (R2 = 0.51–0.74; NRMSE = 8–12%). In contrast, carbon and lignin consistently exhibited lower predictive performance (R2 = 0.32–0.59; NRMSE = 9–15%). Despite covering only the VNIR spectral range, the UAV-based Nano imagery achieved accuracies comparable to those obtained using the airborne full-spectrum Hyspex imagery, indicating that high spatial resolution can partially compensate for limited spectral coverage by reducing soil background effects. The VNIR spectral region alone provided trait estimation accuracies comparable to those obtained using the full visible-to-shortwave infrared (VSWIR) spectrum, whereas SWIR wavelengths contributed only marginal improvements for a subset of structural traits. Among preprocessing approaches, vector normalization generally improved prediction performance by reducing the confounding effects of canopy structure and illumination/viewing geometry, whereas NIRv-adjusted spectra provided limited benefits. Our findings demonstrate that UAV-based VNIR imaging spectroscopy can provide accurate and cost-effective estimation of grassland foliar functional traits. The results also highlight important trade-offs between spectral and spatial resolution in hyperspectral remote sensing and provide practical guidance for selecting imaging spectroscopy platforms and preprocessing approaches for grassland ecosystem monitoring.

Human activities are rapidly transforming ecosystems worldwide, underscoring the urgent need for scalable approaches to monitor vegetation condition and recovery at scale. Public satellite missions such as Sentinel-2 have great potential for monitoring key vegetation biophysical properties because of their high spatial and temporal resolution. Radiative transfer models provide a physically based link between reflectance and vegetation structure and function, and radiative transfer-based inversion methods are well established for retrieving biophysical variables. However, existing operational products are either coarse in resolution, geographically restricted, or available only through on-demand processing workflows that require technical expertise. As a result, access to high-resolution, analysis-ready vegetation biophysical information remains limited, constraining large-scale ecosystem monitoring and restoration assessment. Here, we introduce S2BIOPHYS, the first global operational product providing precomputed, annual maps at high resolution (20 m) for three key vegetation biophysical properties: effective leaf area index (LAIe), fraction of absorbed photosynthetically active radiation (FAPAR) and fractional vegetation cover (FCOVER). To generate these products, we systematically extend radiative transfer-based inversion approaches by coupling simulation-driven model training with empirically constrained parameter optimization using an extensive in-situ reference dataset. Annual composites for 2019-2025 are derived from peak growing-season Sentinel-2 observations and include per-pixel mean values, calibrated uncertainty estimates and observation counts. Validation against more than 11,000 ground reference measurements, evaluated using spatially independent cross-validation, demonstrates strong performance relative to existing retrieval approaches. The resulting ~20 TB global dataset, together with an accompanying Python package enabling sub-annual retrievals, supports downstream applications in biodiversity assessment, ecosystem restoration monitoring, and global reporting frameworks.

Recent advances in remote sensing of biodiversity have resulted in reliable and reproducible estimates of plant traits and biomass, especially when using imaging spectroscopy (IS) data, yet multi-temporal applications remain limited due to restricted data availability and challenges arising from varying acquisition conditions across time steps. In this study, we examined the effect of topographic correction methods on yearly biomass predictions in alpine grasslands using multi-temporal IS data over the Swiss National Park (SNP). We used 2010, 2012, and 2013 Airborne Prism EXperiment (APEX) data together with Analytical Spectral Devices (ASD) field spectrometer data to compare the effects of three topographic correction methods, i.e., (i) Lambertian, (ii) Minnaert, and (iii) modified Minnaert using partial least squares regression (PLSR) models. We found that pair-wise reflectance differences between the correction methods ranged from 0.1 to 8.4% per wavelength per year. Reflectance from both Minnaert methods aligned more closely with that from the handheld spectrometer than reflectance from the Lambertian method. We also found that the biomass models accounted for 61-88% of the variance in calibration, with higher performance for models using Minnaert methods corrected data in 2010 and 2013, and for the Lambertian method corrected data in 2012. Either of the Minnaert methods corrected data generally outperformed that of the Lambertian method, as the latter tended to overcorrect on steep slopes. Differences among methods and years could be associated with variation in slope, aspect, and canopy development stages, with the 2012 data collected in the later season under higher accumulated growing degree days (AGDD). These results imply that temporal comparisons are associated with several constraints and that these need to be addressed when interpreting repeated biomass estimates.

Grassland aboveground biomass provides key insights into ecological processes such as carbon sequestration, animal movement patterns, and agricultural management practices. Different model types have been developed to estimate grassland biomass from satellite imagery. However, differences in model performance across sites with varying management and ecology remain largely understudied. In this study, we compared accuracy and transferability of empirical, physically-based, and hybrid models to estimate grassland biomass from multispectral Sentinel-2 data in an agnostic scenario, i.e., the models were not provided with any site-specific information beyond the spectral data. Based on field data from five study sites in Europe and the United States, we assessed (1)site-level accuracy of biomass estimation models, (2) model transferability between sites (domain shift), (3) the performance of models trained or optimized with data from multiple study sites (domain generalization), and (4) the relationship between epistemic uncertainty and model transferability. Our results showed that (1) all models exhibited comparable performance at the site level, (2) physically-based models showed the highest degree of transferability between sites, (3) no model consistently outperformed all other models when trained or optimized with field data from multiple sites, and (4) epistemic uncertainty was not necessarily a reliable measure of model applicability to unseen data. Our findings demonstrate the challenges associated with grassland biomass models under domain shift. This elucidates limits to agnostic inference in targeting diverse grasslands and highlights that model transferability is an integral part of performance assessment towards scalable satellite-based grassland monitoring systems, especially as the community increasingly deploys models at continental to global scales.

Assessing plant diversity using remote sensing, including airborne imaging spectroscopy, shows promise for large-scale biodiversity monitoring in landscape restoration and conservation. Enriching plantations with native trees is a key restoration strategy to enhance biodiversity and ecosystem functions in agricultural lands. In this study, we tested how well imaging spectroscopy characterizes plant diversity in 37 experimental plots of varying sizes and planted diversity levels in a biodiversity-enriched oil palm plantation in Sumatra, Indonesia. Six years after establishing the plots, we acquired airborne imaging spectroscopy data comprising 160 spectral bands (400-1000 nm, at similar to 3.7 nm bandwidth) at 0.3 m spatial resolution. We calculated spectral diversity as the variance among image pixels and partitioned spectral diversity into alpha and beta diversity components. After controlling for differences in sampling area through rarefaction, we found no significant relationship between spectral and plant alpha diversity. Further, the relationships between the local contribution of spectral beta diversity and plant beta diversity revealed no significant trends. Spectral variability within plots was substantially higher than among plots (spectral alpha diversity similar to 82%-87%, spectral beta diversity similar to 11%-18%). These discrepancies are likely due to the structural dominance of oil palm crowns, which absorbed most of the light, while most of the plant diversity occurring below the oil palm canopy was not detectable by airborne spectroscopy. Our study highlights that remote sensing of plant diversity in ecosystems with strong vertical stratification and high understory diversity, such as agroforests, would benefit from combining data from passive with data from active sensors, such as LiDAR, to capture structural diversity.

Tracking biodiversity across biomes over space and time has emerged as an imperative in unified global efforts to manage our living planet for a sustainable future for humanity. We harness the National Ecological Observatory Network to develop routines using airborne spectroscopic imagery to predict multiple dimensions of plant biodiversity at continental scale across biomes in the US. Our findings show strong and positive associations between diversity metrics based on spectral species and ground-based plant species richness and other dimensions of plant diversity, whereas metrics based on distance matrices did not. We found that spectral diversity consistently predicts analogous metrics of plant taxonomic, functional, and phylogenetic dimensions of biodiversity across biomes. The approach demonstrates promise for monitoring dimensions of biodiversity globally by integrating ground-based measures of biodiversity with imaging spectroscopy and advances capacity toward a Global Biodiversity Observing System.

Global change is associated with variable shifts in the annual production of aboveground plant biomass, suggesting localized sensitivities with unclear causal origins. Combining remotely sensed normalized difference vegetation index data since the 1980s with contemporary field data from 84 grasslands on 6 continents, we show a widening divergence in site-level biomass ranging from +51% to -34% globally. Biomass generally increased in warmer, wetter and species-rich sites with longer growing seasons and declined in species-poor arid areas. Phenological changes were widespread, revealing substantive transitions in grassland seasonal cycling. Grazing, nitrogen deposition and plant invasion were prevalent in some regions but did not predict overall trends. Grasslands are undergoing sizable changes in production, with implications for food security, biodiversity and carbon storage especially in arid regions where declines are accelerating. Remotely sensed NDVI data and contemporary field data from 84 grasslands on 6 continents show increasing divergence in aboveground plant biomass between sites in different bioclimatic regions.

Characterizing foliar trait variation in sun and shade leaves can provide insights into inter- and intra-species resource use strategies and plant response to environmental change. However, datasets with records of multiple foliar traits from the same individual and including shade leaves are sparse, which limits our ability to investigate trait-trait, trait-environment relationships and trait coordination in both sun and shade leaves. We presented a comprehensive dataset of 15 foliar traits from sun and shade leaves sampled with leaf spectroscopy, including 424 individuals of 110 plant species from 19 sites across eastern North America. We investigated trait variation, covariation, scaling relationships with leaf mass, and the effects of environment, canopy position, and taxonomy on trait expression. Generally, sun leaves had higher leaf mass per area, nonstructural carbohydrates and total phenolics, lower mass-based chlorophyll a + b, carotenoids, phosphorus, and potassium, but exhibited species-specific characteristics. Covariation between sun and shade leaf traits, and trait-environment relationships were overall consistent across species. The main dimensions of foliar trait variation in seed plants were revealed including leaf economics traits, photosynthetic pigments, defense, and structural traits. Taxonomy and canopy position collectively explained most of the foliar trait variation. This study highlights the importance of including intra-individual and intra-specific trait variation to improve our understanding of ecosystem functions. Our findings have implications for efficient field sampling, and trait mapping with remote sensing.

Standard and easily accessible cross-thematic spatial databases are key resources in ecological research. In Switzerland, as in many other countries, available data are scattered across computer servers of research institutions and are rarely provided in standard formats (e.g., different extents or projections systems, inconsistent naming conventions). Consequently, their joint use can require heavy data management and geomatic operations. Here, we introduce SWECO25, a Swiss-wide raster database at 25-meter resolution gathering 5,265 layers. The 10 environmental categories included in SWECO25 are: geologic, topographic, bioclimatic, hydrologic, edaphic, land use and cover, population, transportation, vegetation, and remote sensing. SWECO25 layers were standardized to a common grid sharing the same resolution, extent, and geographic coordinate system. SWECO25 includes the standardized source data and newly calculated layers, such as those obtained by computing focal or distance statistics. SWECO25 layers were validated by a data integrity check, and we verified that the standardization procedure had a negligible effect on the output values. SWECO25 is available on Zenodo and is intended to be updated and extended regularly.

Leaf spectra are integrated foliar phenotypes that capture a range of traits and can provide insight into ecological processes. Leaf traits, and therefore leaf spectra, may reflect belowground processes such as mycorrhizal associations. However, evidence for the relationship between leaf traits and mycorrhizal association is mixed, and few studies account for shared evolutionary history.We conduct partial least squares discriminant analysis to assess the ability of spectra to predict mycorrhizal type. We model the evolution of leaf spectra for 92 vascular plant species and use phylogenetic comparative methods to assess differences in spectral properties between arbuscular mycorrhizal and ectomycorrhizal plant species.Partial least squares discriminant analysis classified spectra by mycorrhizal type with 90% (arbuscular) and 85% (ectomycorrhizal) accuracy. Univariate models of principal components identified multiple spectral optima corresponding with mycorrhizal type due to the close relationship between mycorrhizal type and phylogeny. Importantly, we found that spectra of arbuscular mycorrhizal and ectomycorrhizal species do not statistically differ from each other after accounting for phylogeny.While mycorrhizal type can be predicted from spectra, enabling the use of spectra to identify belowground traits using remote sensing, this is due to evolutionary history and not because of fundamental differences in leaf spectra due to mycorrhizal type.

Summary Plant ecologists use functional traits to describe how plants respond to and influence their environment. Reflectance spectroscopy can provide rapid, non‐destructive estimates of leaf traits, but it remains unclear whether general trait‐spectra models can yield accurate estimates across functional groups and ecosystems. We measured leaf spectra and 22 structural and chemical traits for nearly 2000 samples from 103 species. These samples span a large share of known trait variation and represent several functional groups and ecosystems, mainly in eastern Canada. We used partial least‐squares regression (PLSR) to build empirical models for estimating traits from spectra. Within the dataset, our PLSR models predicted traits such as leaf mass per area (LMA) and leaf dry matter content (LDMC) with high accuracy (R2 > 0.85; %RMSE < 10). Models for most chemical traits, including pigments, carbon fractions, and major nutrients, showed intermediate accuracy (R2 = 0.55–0.85; %RMSE = 12.7–19.1). Micronutrients such as Cu and Fe showed the poorest accuracy. In validation on external datasets, models for traits such as LMA and LDMC performed relatively well, while carbon fractions showed steep declines in accuracy. We provide models that produce fast, reliable estimates of several functional traits from leaf spectra. Our results reinforce the potential uses of spectroscopy in monitoring plant function around the world.

1.Selecting the best subset of covariates out of a panel of many candidates is a key and highly influential stage of the species distribution modelling process. Yet, there is currently no commonly accepted and widely adopted standard approach by which to perform this selection.2.We introduce a two-step “embedded” covariate selection procedure aimed at optimizing the predictive ability and parsimony of species distribution models fitted in a context of high-dimensional candidate covariate space. The procedure combines a collinearity-filtering algorithm (Step A) with three model-specific embedded regularization techniques (Step B), including generalized linear model with elastic net regularization, generalized additive model with null-space penalization, and guided regularized random forest.3.We evaluated the embedded covariate selection procedure through an example application aimed at modelling the habitat suitability of 50 species in Switzerland from a suite of 123 candidate covariates. We demonstrated the ability of the embedded covariate selection procedure to provide significantly more accurate species distribution models as compared to models obtained with alternative procedures. Model performance was independent of the characteristics of the species data, such as the number of occurrence records or their spatial distribution across the study area.4.We implemented and streamlined our embedded covariate selection procedure in the covsel R package, paving the way for a ready-to-use, automated, covariate selection tool that was missing in the field of species distribution modelling. All the information required for installing and running the covsel R package is openly available on the GitHub repository https://github.com/N-SDM/covsel.

&lt;p&gt;Biodiversity science is entering a new era. Years of effort by the scientific community are culminating in recent and upcoming launches of satellite systems specifically designed for global biodiversity assessment and monitoring. In addition, the Kunming-Montr&amp;#233;al Global Biodiversity Framework, the most ambitious international agreement addressing biodiversity loss, is pushing for biodiversity and ecosystem protection, restoration and better management to occupy more prominent positions on global political agendas. It is clear by now that we need remote sensing to assess the status and monitor biodiversity globally and repeatedly. However, global biodiversity observatory systems need to combine remote sensing with ground observations to develop reliable and intepretable products. In this talk, I will try to summarize the current status and potential future directions of remote sensing of biodiversity across spatial, temporal and biological scales. The focus of my talk will be plant spectroscopy, which is based on the physical and physiological connections between plants and light. I will discuss the ways in which integrating ecological theory with measurements across spatial and temporal scales allow for a better understanding of what aspects of biodiversity global satellite systems are capable of detecting on the ground. I will also provide examples of remote sensing studies investigating the diversity of taxonomic groups other than plants through their connection with particular vegetation characteristics. Future advances in the field of remote sensing of biodiversity will benefit more than ever from diverse teams, global cooperation and collaborations across disciplines, including biology, geography, computer science and robotics. Now is the time to do our best work to help prevent and mitigate the negative consequences of biodiversity loss.&lt;/p&gt;

&lt;p&gt;Accurate retrieval of biophysical variables is crucial for characterizing properties (i.e. traits) of plant canopies and capturing their spatiotemporal changes. Optical remote sensing offers the unique possibility for frequent and large-scale mapping of biophysical variables due to strong associations between spectral data and plant optical traits. One approach to formally predict plant properties remotely is based on hybrid retrieval. In this approach, a radiative transfer model (RTM) is used to simulate plant spectra for model training and then a machine learning regression is utilized for model prediction.&lt;/p&gt; &lt;p&gt;Hybrid retrieval approaches have two main advantages. First, the approach augments large field datasets needed for training with simulations modelled based on physical relationships between electromagnetic radiation and plant properties. The universal physics behind this have led to assumptions of greater transferability of these models when compared to empirical models. Second, the machine learning implementation provides the flexibility and computational efficiency of nonlinear nonparametric methods to link spectra and plant properties.&lt;/p&gt; &lt;p&gt;The recent implementation of active learning (AL) approaches offers promising and adaptive solutions to further enhance hybrid retrieval approaches. AL seeks to overcome the genericity and heavy assumptions of RTM simulations as opposed to the noisy real-world spectra and particularities of ecosystems by subsetting the training data to boost model performance. However, it is unclear how the selection of training data by an AL approach thereby affects model transferability and whether its selection relates to the ecology of different sites. Our work aims to assess how representative the AL-selected training samples are for their respective ecosystem and whether the generated models are transferable to other study sites.&lt;/p&gt; &lt;p&gt;Here, we used Gaussian process regression (GPR) trained with PROSAIL simulations in combination with AL to retrieve canopy foliar biomass and nitrogen content from Sentinel-2 data in three grassland sites with different characteristics, including alpine, prairie, and temperate grasslands in Switzerland, the United States, and Germany, respectively, and one heterogeneous forest and shrubland site in Portugal. We compared the trait space of the selected training samples with those of in-situ data and TRY database to assess their respective ecological representativeness. Further, we used our generated models to predict canopy foliar biomass and nitrogen across sites to check for their transferability.&lt;/p&gt; &lt;p&gt;Our preliminary results show promising accuracy of locally trained models to retrieve canopy foliar biomass (Switzerland: R&lt;sup&gt;2&lt;/sup&gt; = 0.41, RMSE = 106.5 g/m&lt;sup&gt;2&lt;/sup&gt;; United States: R&lt;sup&gt;2&lt;/sup&gt; = 0.42, RMSE = 85.5 g/m&lt;sup&gt;2&lt;/sup&gt;; Germany: R&lt;sup&gt;2&lt;/sup&gt; = 0.28, RMSE = 96.2 g/m&lt;sup&gt;2&lt;/sup&gt;; Portugal: R&lt;sup&gt;2&lt;/sup&gt; = 0.6, RMSE = 60.9 g/m&lt;sup&gt;2&lt;/sup&gt;). In particular, AL-selected training data increased model performances but was highly affected by the validation data thus limiting the general transferability of the models across study sites.&lt;/p&gt; &lt;p&gt;Based on these results, we can confirm adequate and stable performance of locally trained GPR-AL models. However, the transferability of such an approach requires further testing and an expanded search for solutions. For now, strong trade-offs exist between local optimization and transferability which challenges predictions of high accuracy across large spatial extents with limited field data.&lt;/p&gt;

Spectroscopy at the leaf and canopy scales has attracted considerable interest in plant ecology over the past decades. Using reflectance spectra, ecologists can infer plant traits and strategies—and the community‐ or ecosystem‐level processes they correlate with—at individual or community levels, covering more individuals and larger areas than traditional field surveys. Because of the complex entanglement of structural and chemical factors that generate spectra, it can be tricky to understand exactly what phenotypic information they contain. We discuss common approaches to estimating plant traits from spectra—radiative transfer and empirical models—and elaborate on their strengths and limitations in terms of the causal influences of various traits on the spectrum. Many chemical traits have broad, shallow and overlapping absorption features, and we suggest that covariance among traits may have an important role in giving empirical models the flexibility to estimate such traits. While trait estimates from reflectance spectra have been used to test ecological hypotheses over the past decades, there is also a growing body of research that uses spectra directly, without estimating specific traits. By treating positions of species in multidimensional spectral space as analogous to trait space, researchers can infer processes that structure plant communities using the information content of the full spectrum, which may be greater than any standard set of traits. We illustrate this power by showing that co‐occurring grassland species are more separable in spectral space than in trait space and that the intrinsic dimensionality of spectral data is comparable to fairly comprehensive trait datasets. Nevertheless, using spectra this way may make it harder to interpret patterns in terms of specific biological processes. Synthesis . Plant spectra integrate many aspects of plant form and function. The information in the spectrum can be distilled into estimates of specific traits, or the spectrum can be used in its own right. These two approaches may be complementary—the former being most useful when specific traits of interest are known in advance and reliable models exist to estimate them, and the latter being most useful under uncertainty about which aspects of function matter most.

Forest ecosystems sequester approximately half of the world’s organic carbon (C), most of it in the soil. The amount of soil C stored depends on the input and decomposition rate of soil organic matter (OM), which is controlled by the abundance and composition of the microbial and invertebrate communities, soil physico-chemical properties, and (micro)-climatic conditions. Although many studies have assessed how these site-specific climatic and soil properties affect the decomposition of fresh OM, differences in the type and quality of the OM substrate used, make it difficult to compare and extrapolate results across larger scales. Here, we used standard wood stakes made from aspen (Populus tremuloides Michx.) and loblolly pine (Pinus taeda L.) to explore how climate and abiotic soil properties affect wood decomposition across 44 unharvested forest stands located across the northern hemisphere. Stakes were placed in three locations: (i) on top of the surface organic horizons (surface), (ii) at the interface between the surface organic horizons and mineral soil (interface), and (iii) into the mineral soil (mineral). Decomposition rates of both wood species was greatest for mineral stakes and lowest for stakes placed on the surface organic horizons, but aspen stakes decomposed faster than pine stakes. Our models explained 44 and 36% of the total variation in decomposition for aspen surface and interface stakes, but only 0.1% (surface), 12% (interface), 7% (mineral) for pine, and 7% for mineral aspen stakes. Generally, air temperature was positively, precipitation negatively related to wood stake decomposition. Climatic variables were stronger predictors of decomposition than soil properties (surface C:nitrogen ratio, mineral C concentration, and pH), regardless of stake location or wood species. However, climate-only models failed in explaining wood decomposition, pointing toward the importance of including local-site properties when predicting wood decomposition. The difficulties we had in explaining the variability in wood decomposition, especially for pine and mineral soil stakes, highlight the need to continue assessing drivers of decomposition across large global scales to better understand and estimate surface and belowground C cycling, and understand the drivers and mechanisms that affect C pools, CO2 emissions, and nutrient cycles.

Monitoring the rapid and extensive changes in plant species distributions occurring worldwide requires large-scale, continuous and repeated biodiversity assessments. Imaging spectrometers are at the core of novel spaceborne sensor fleets designed for this task, but the degree to which they can capture plant species composition and diversity across ecosystems has yet to be determined. Here we use imaging spectroscopy and vegetation data collected by the National Ecological Observatory Network (NEON) to show that at the landscape level, spectral beta-diversity-calculated directly from spectral images-captures changes in plant species composition across all major biomes in the United States ranging from arctic tundra to tropical forests. At the local level, however, the relationship between spectral alpha- and plant alpha-diversity was positive only at sites with high canopy density and large plant-to-pixel size. Our study demonstrates that changes in plant species composition and diversity can be effectively and reliably assessed with imaging spectroscopy across terrestrial ecosystems at the beta-diversity scale-the spatial scale of spaceborne missions-paving the way for close-to-real-time biodiversity monitoring at the planetary level.

The oak wilt disease caused by the invasive fungal pathogen Bretziella fagacearum is one of the greatest threats to oak-dominated forests across the Eastern United States. Accurate detection and monitoring over large areas are necessary for management activities to effectively mitigate and prevent the spread of oak wilt. Canopy spectral reflectance contains both phylogenetic and physiological information across the visible near-infrared (VNIR) and short-wave infrared (SWIR) ranges that can be used to identify diseased red oaks. We develop partial least square discriminant analysis (PLS-DA) models using airborne hyperspectral reflectance to detect diseased canopies and assess the importance of VNIR, SWIR, phylogeny, and physiology for oak wilt detection. We achieve high accuracy through a three-step phylogenetic process in which we first distinguish oaks from other species (90% accuracy), then red oaks from white oaks ( Quercus macrocarpa ) (93% accuracy), and, lastly, infected from non-infected trees (80% accuracy). Including SWIR wavelengths increased model accuracy by ca. 20% relative to models based on VIS-NIR wavelengths alone; using a phylogenetic approach also increased model accuracy by ca. 20% over a single-step classification. SWIR wavelengths include spectral information important in differentiating red oaks from other species and in distinguishing diseased red oaks from healthy red oaks. We determined the most important wavelengths to identify oak species, red oaks, and diseased red oaks. We also demonstrated that several multispectral indices associated with physiological decline can detect differences between healthy and diseased trees. The wavelengths in these indices also tended to be among the most important wavelengths for disease detection within PLS-DA models, indicating a convergence of the methods. Indices were most significant for detecting oak wilt during late August, especially those associated with canopy photosynthetic activity and water status. Our study suggests that coupling phylogenetics, physiology, and canopy spectral reflectance provides an interdisciplinary and comprehensive approach that enables detection of forest diseases at large scales. These results have potential for direct application by forest managers for detection to initiate actions to mitigate the disease and prevent pathogen spread.

A core goal of the National Ecological Observatory Network (NEON) is to measure changes in biodiversity across the 30-yr horizon of the network. In contrast to NEON's extensive use of automated instruments to collect environmental data, NEON's biodiversity surveys are almost entirely conducted using traditional human-centric field methods. We believe that the combination of instrumentation for remote data collection and machine learning models to process such data represents an important opportunity for NEON to expand the scope, scale, and usability of its biodiversity data collection while potentially reducing long-term costs. In this manuscript, we first review the current status of instrument-based biodiversity surveys within the NEON project and previous research at the intersection of biodiversity, instrumentation, and machine learning at NEON sites. We then survey methods that have been developed at other locations but could potentially be employed at NEON sites in future. Finally, we expand on these ideas in five case studies that we believe suggest particularly fruitful future paths for automated biodiversity measurement at NEON sites: acoustic recorders for sound-producing taxa, camera traps for medium and large mammals, hydroacoustic and remote imagery for aquatic diversity, expanded remote and ground-based measurements for plant biodiversity, and laboratory-based imaging for physical specimens and samples in the NEON biorepository. Through its data science-literate staff and user community, NEON has a unique role to play in supporting the growth of such automated biodiversity survey methods, as well as demonstrating their ability to help answer key ecological questions that cannot be answered at the more limited spatiotemporal scales of human-driven surveys.