Search

Measuring forest biodiversity using terrestrial surveys is expensive and can only capture common species abundance in large heterogeneous landscapes. In contrast, combining airborne imagery with computer vision can generate individual tree data at the scales of hundreds of thousands of trees. To train computer vision models, ground-based species labels are combined with airborne reflectance data. Due to the difficulty of finding rare species in a large landscape, the majority of classification models only include the most abundant species, leading to biased predictions at broad scales. Extending classification models to include rare species requires targeted data collection and algorithmic improvements to overcome large data imbalances between dominant and rare taxa. In addition, large landscapes often require multiple acquisition events, leading to significant within-species variation in reflectance spectra. Using a multi-temporal hierarchical model, we demonstrate the ability to include species predicted at less than 1% frequency in landscape without losing performance on the dominant species. The final model has over 75% accuracy for 14 species with improved rare species classification compared to a baseline deep learning model. After filtering out dead trees, we generate landscape species maps of individual crowns for over 670,000 individual trees at the Ordway Swisher Biological Station within the National Ecological Observatory Network. We estimate the relative abundance of the species within the landscape and provide three measures of uncertainty to generate a range of counts for each species. These maps provide the first estimates of canopy tree diversity within NEON sites to include rare species and provide a blueprint for capturing tree diversity using airborne computer vision at broad scales.

Advances in artificial intelligence for computer vision hold great promise for increasing the scales at which ecological systems can be studied. The distribution and behavior of individuals is central to ecology, and computer vision using deep neural networks can learn to detect individual objects in imagery. However, developing supervised models for ecological monitoring is challenging because it requires large amounts of human-labeled training data, requires advanced technical expertise and computational infrastructure, and is prone to overfitting. This limits application across space and time. One solution is developing generalized models that can be applied across species and ecosystems. Using over 250,000 annotations from 13 projects from around the world, we develop a general bird detection model that achieves over 65% recall and 50% precision on novel aerial data without any local training despite differences in species, habitat, and imaging methodology. Fine-tuning this model with only 1000 local annotations increases these values to an average of 84% recall and 69% precision by building on the general features learned from other data sources. Retraining from the general model improves local predictions even when moderately large annotation sets are available and makes model training faster and more stable. Our results demonstrate that general models for detecting broad classes of organisms using airborne imagery are achievable. These models can reduce the effort, expertise, and computational resources necessary for automating the detection of individual organisms across large scales, helping to transform the scale of data collection in ecology and the questions that can be addressed.

Scene understanding methodologies (e.g., classification, segmentation, and anomaly detection) can be costly when operating on a per-pixel basis. As remote sensing applications rely on imagery with ...

Advances in remote sensing imagery and machine learning applications unlock the potential for developing algorithms for species classification at the level of individual tree crowns at unprecedented scales. However, most approaches to date focus on site-specific applications and a small number of taxonomic groups. Little is known about how well these approaches generalize across broader geographic areas and ecosystems. Leveraging field surveys and hyperspectral remote sensing data from the National Ecological Observatory Network (NEON), we developed a continental-extent model for tree species classification that can be applied to the network, including a wide range of US terrestrial ecosystems. We compared the performance of a model trained with data from 27 NEON sites to models trained with data from each individual site, evaluating advantages and challenges posed by training species classifiers at the US scale. We evaluated the effect of geographic location, topography, and ecological conditions on the accuracy and precision of species predictions (72 out of 77 species). On average, the general model resulted in good overall classification accuracy (micro-F1 score), with better accuracy than site-specific classifiers (average individual tree level accuracy of 0.77 for the general model and 0.70 for site-specific models). Aggregating species to the genus-level increased accuracy to 0.83. Regions with more species exhibited lower classification accuracy. Predicted species were more likely to be confused with congeneric and co-occurring species and confusion was highest for trees with structural damage and in complex closed-canopy forests. The model produced accurate estimates of uncertainty, correctly identifying trees where confusion was likely. Using only data from NEON, this single integrated classifier can make predictions for 20% of all tree species found in forest ecosystems across the entire US, which make up to roughly 90% of the upper canopy of the studied ecosystems. This suggests the potential for integrating information from multiple datasets and locations to develop broad scale general models for species classification from hyperspectral imaging.

Delineating and classifying individual trees in remote sensing data is challenging. Many tree crown delineation methods have difficulty in closed-canopy forests and do not leverage multiple datasets. Methods to classify individual species are often accurate for common species, but perform poorly for less common species and when applied to new sites. We ran a data science competition to help identify effective methods for delineation of individual crowns and classification to determine species identity. This competition included data from multiple sites to assess the methods’ ability to generalize learning across multiple sites simultaneously, and transfer learning to novel sites where the methods were not trained. Six teams, representing 4 countries and 9 individual participants, submitted predictions. Methods from a previous competition were also applied and used as the baseline to understand whether the methods are changing and improving over time. The best delineation method was based on an instance segmentation pipeline, closely followed by a Faster R-CNN pipeline, both of which outperformed the baseline method. However, the baseline (based on a growing region algorithm) still performed well as did the Faster R-CNN. All delineation methods generalized well and transferred to novel forests effectively. The best species classification method was based on a two-stage fully connected neural network, which significantly outperformed the baseline (a random forest and Gradient boosting ensemble). The classification methods generalized well, with all teams training their models using multiple sites simultaneously, but the predictions from these trained models generally failed to transfer effectively to a novel site. Classification performance was strongly influenced by the number of field-based species IDs available for training the models, with most methods predicting common species well at the training sites. Classification errors (i.e., species misidentification) were most common between similar species in the same genus and different species that occur in the same habitat. The best methods handled class imbalance well and learned unique spectral features even with limited data. Most methods performed better than baseline in detecting new (untrained) species, especially in the site with no training data. Our experience further shows that data science competitions are useful for comparing different methods through the use of a standardized dataset and set of evaluation criteria, which highlights promising approaches and common challenges, and therefore advances the ecological and remote sensing field as a whole.

Advances in artificial intelligence for computer vision hold great promise for increasing the scales at which ecological systems can be studied. The distribution and behavior of individuals is central to ecology, and computer vision using deep neural networks can learn to detect individual objects in imagery. However, developing supervised models for ecological monitoring is challenging because it needs large amounts of human-labeled training data, requires advanced technical expertise and computational infrastructure, and is prone to overfitting. This limits application across space and time. One solution is developing generalized models that can be applied across species and ecosystems. Using over 250,000 annotations from 13 projects from around the world, we develop a general bird detection model that achieves over 65% recall and 50% precision on novel aerial data without any local training despite differences in species, habitat, and imaging methodology. Fine-tuning this model with only 1000 local annotations increase these values to an average of 84% recall and 69% precision by building on the general features learned from other data sources. Retraining from the general model improves local predictions even when moderately large annotation sets are available and makes model training faster and more stable. Our results demonstrate that general models for detecting broad classes of organisms using airborne imagery are achievable. These models can reduce the effort, expertise, and computational resources necessary for automating the detection of individual organisms across large scales, helping to transform the scale of data collection in ecology and the questions that can be addressed.

Broad scale remote sensing promises to build forest inventories at unprecedented scales. A crucial step in this process is to associate sensor data into individual crowns. While dozens of crown detection algorithms have been proposed, their performance is typically not compared based on standard data or evaluation metrics. There is a need for a benchmark dataset to minimize differences in reported results as well as support evaluation of algorithms across a broad range of forest types. Combining RGB, LiDAR and hyperspectral sensor data from the USA National Ecological Observatory Network’s Airborne Observation Platform with multiple types of evaluation data, we created a benchmark dataset to assess crown detection and delineation methods for canopy trees covering dominant forest types in the United States. This benchmark dataset includes an R package to standardize evaluation metrics and simplify comparisons between methods. The benchmark dataset contains over 6,000 image-annotated crowns, 400 field-annotated crowns, and 3,000 canopy stem points from a wide range of forest types. In addition, we include over 10,000 training crowns for optional use. We discuss the different evaluation data sources and assess the accuracy of the image-annotated crowns by comparing annotations among multiple annotators as well as overlapping field-annotated crowns. We provide an example submission and score for an open-source algorithm that can serve as a baseline for future methods., Author summary Combining RGB, LiDAR and hyperspectral sensor data from the USA National Ecological Observatory Network’s Airborne Observation Platform with multiple types of evaluation data, we created a benchmark dataset to assess crown detection and delineation methods for canopy trees covering dominant forest types in the United States. This benchmark dataset includes an R package to standardize evaluation metrics and simplify comparisons between methods. The benchmark dataset contains over 6,000 image-annotated crowns, 400 field-annotated crowns, and 3,000 canopy stem points from a wide range of forest types.

Supervised methods for object delineation in remote sensing require labeled ground-truth data. Gathering sufficient high quality ground-truth data is difficult, especially when targets are of irregular shape or difficult to distinguish from background or neighboring objects. Tree crown delineation provides key information from remote sensing images for forestry, ecology, and management. However, tree crowns in remote sensing imagery are often difficult to label and annotate due to irregular shape, overlapping canopies, shadowing, and indistinct edges. There are also multiple approaches to annotation in this field (e.g., rectangular boxes vs. convex polygons) that further contribute to annotation imprecision. However, current evaluation methods do not account for this uncertainty in annotations, and quantitative metrics for evaluation can vary across multiple annotators. In this paper, we address these limitations by developing an adaptation of the Rand index for weakly-labeled crown delineation that we call RandCrowns. Our new RandCrowns evaluation metric provides a method to appropriately evaluate delineated tree crowns while taking into account imprecision in the ground-truth delineations. The RandCrowns metric reformulates the Rand index by adjusting the areas over which each term of the index is computed to account for uncertain and imprecise object delineation labels. Quantitative comparisons to the commonly used intersection over union method shows a decrease in the variance generated by differences among multiple annotators. Combined with qualitative examples, our results suggest that the RandCrowns metric is more robust for scoring target delineations in the presence of uncertainty and imprecision in annotations that are inherent to tree crown delineation.

Broad scale remote sensing promises to build forest inventories at unprecedented scales. A crucial step in this process is designing individual tree segmentation algorithms to associate pixels into delineated tree crowns. While dozens of tree delineation algorithms have been proposed, their performance is typically not compared based on standard data or evaluation metrics, making it difficult to understand which algorithms perform best under what circumstances. There is a need for an open evaluation benchmark to minimize differences in reported results due to data quality, forest type and evaluation metrics, and to support evaluation of algorithms across a broad range of forest types. Combining RGB, LiDAR and hyperspectral sensor data from the National Ecological Observatory Network’s Airborne Observation Platform with multiple types of evaluation data, we created a novel benchmark dataset to assess individual tree delineation methods. This benchmark dataset includes an R package to standardize evaluation metrics and simplify comparisons between methods. The benchmark dataset contains over 6,000 image-annotated crowns, 424 field-annotated crowns, and 3,777 overstory stem points from a wide range of forest types. In addition, we include over 10,000 training crowns for optional use. We discuss the different evaluation sources and assess the accuracy of the image-annotated crowns by comparing annotations among multiple annotators as well as to overlapping field-annotated crowns. We provide an example submission and score for an open-source baseline for future methods.

Forests provide essential biodiversity, ecosystem and economic services. Information on individual trees is important for understanding the state of forest ecosystems but obtaining individual-level data at broad scales is challenging due to the costs and logistics of data collection. While advances in remote sensing techniques allow surveys of individual trees at unprecedented extents, there remain significant technical and computational challenges in turning sensor data into tangible information. Using deep learning methods, we produced an open-source dataset of individual-level crown estimates for 100 million trees at 37 sites across the United States surveyed by the National Ecological Observatory Network’s Airborne Observation Platform. Each canopy tree crown is represented by a rectangular bounding box and includes information on the height, crown area, and spatial location of the tree. Tree crowns identified using this technique correspond well with hand-labeled crowns, exhibiting both high levels of overlap and good correspondence in height estimates. These data have the potential to drive significant expansion of individual-level research on trees by facilitating both regional analyses at scales of ~10,000 ha and cross-region comparisons encompassing forest types from most of the United States.

Functional ecology has increasingly focused on describing ecological communities based on their traits (measurable features affecting individuals' fitness and performance). Analyzing trait distributions within and among forests could significantly improve understanding of community composition and ecosystem function. Historically, data on trait distributions are generated by (1) collecting a small number of leaves from a small number of trees, which suffers from limited sampling but produces information at the fundamental ecological unit (the individual), or (2) using remote-sensing images to infer traits, producing information continuously across large regions, but as plots (containing multiple trees of different species) or pixels, not individuals. Remote-sensing methods that identify individual trees and estimate their traits would provide the benefits of both approaches, producing continuous large-scale data linked to biological individuals. We used data from the National Ecological Observatory Network (NEON) to develop a method to scale up functional traits from 160 trees to the millions of trees within the spatial extent of two NEON sites. The pipeline consists of three stages: (1) image segmentation, to identify individual trees and estimate structural traits; (2) an ensemble of models to infer leaf mass area (LMA), nitrogen, carbon, and phosphorus content using hyperspectral signatures, and DBH from allometry; and (3) predictions for segmented crowns for the full remote-sensing footprint at the NEON sites. The R2 values on held-out test data ranged from 0.41 to 0.75 on held-out test data. The ensemble approach performed better than single partial least-squares models. Carbon performed poorly compared to other traits (R2 of 0.41). The crown segmentation step contributed the most uncertainty in the pipeline, due to over-segmentation. The pipeline produced good estimates of DBH (R2 of 0.62 on held-out data). Trait predictions for crowns performed significantly better than comparable predictions on pixels, resulting in improvement of R2 on test data of between 0.07 and 0.26. We used the pipeline to produce individual-level trait data for ~5 million individual crowns, covering a total extent of ~360 km2 . This large data set allows testing ecological questions on landscape scales, revealing that foliar traits are correlated with structural traits and environmental conditions.

Remote sensing of forested landscapes can transform the speed, scale, and cost of forest research. The delineation of individual trees in remote sensing images is an essential task in forest analysis. Here we introduce a new Python package, DeepForest, that detects individual trees in high resolution RGB imagery using deep learning.While deep learning has proven highly effective in a range of computer vision tasks, it requires large amounts of training data that are typically difficult to obtain in ecological studies. DeepForest overcomes this limitation by including a model pre-trained on over 30 million algorithmically generated crowns from 22 forests and fine-tuned using 10,000 hand-labeled crowns from 6 forests.The package supports the application of this general model to new data, fine tuning the model to new datasets with user labeled crowns, training new models, and evaluating model predictions. This simplifies the process of using and retraining deep learning models for a range of forests, sensors, and spatial resolutions.We illustrate the workflow of DeepForest using data from the National Ecological Observatory Network, a tropical forest in French Guiana, and street trees from Portland, Oregon.

We present a novel design of experiments (DOE) approach to incorporate model identification into optimal experimental designs based on a postulated model superstructure and an associated relaxation strategy. We show that an adaptive online design of experiments allows for the accurate estimation of the parameters of a domain-restricted model, as well as the model structure and domain on which that model is valid. We further show that previous attempts at combining model identification and parameter estimation are a special case of this framework (when the objective function is formulated in terms of the trace of the Fisher information matrix), and thus the proposed formulation provides the option to use alternate or more complex objective functions. The efficacy of the proposed framework is shown through two case studies: a batch reactor with Arrhenius-type reactions and a carbon dioxide adsorption system.

The seasonal movement of animals at broad spatial scales provides insight into life-history, ecology and conservation. By combining high-resolution satellite-tagged data with hierarchical Bayesian movement models, we can associate spatial patterns of movement with marine animal behavior. We used a multi-state mixture model to describe humpback whale traveling and area-restricted search states as they forage along the West Antarctic Peninsula. We estimated the change in the geography, composition and characteristics of these behavioral states through time. We show that whales later in the austral fall spent more time in movements associated with foraging, traveled at lower speeds between foraging areas, and shifted their distribution northward and inshore. Seasonal changes in movement are likely due to a combination of sea ice advance and regional shifts in the primary prey source. Our study is a step towards dynamic movement models in the marine environment at broad scales.

Aim Under climate change, it is likely that as species reshuffle based on their environmental tolerances, novel assemblages will form and some current assemblages will disappear. It is important for future monitoring and conservation that we understand where these novel and disappearing assemblages occur and how they differ among dimensions (taxonomic, phylogenetic and functional) of diversity. Here we investigate the geographical and environmental patterns of novel and disappearing assemblages; whether these patterns hold across dimensions of diversity; and how these assemblages are characterized in trait space. Location Ecuador. Methods We used ensemble species distribution modelling to estimate the distributions of 151 hummingbird species into the projected climate for 2070. Using standard beta diversity measures, we identified novel and disappearing taxonomic, phylogenetic and functional assemblages. Results We found that novel and disappearing hummingbird assemblages are likely under climate change, particularly in extreme environments and with novel assemblages replacing disappearing assemblages. Although the patterns of novel and disappearing assemblages were similar among dimensions of diversity, we found that there were fewest novel and disappearing functional assemblages. The future assemblages were characterized by an increase in functional space, which is counter to typical predictions of trait homogenization under climate change. Main conclusions Novel and disappearing assemblages are likely to pose management challenges for future conservation. Here we present an approach to identify such assemblages. By considering the geographic and environmental context of novel and disappearing assemblages for different dimensions of diversity, we can start to identify the mechanisms behind these patterns.

To determine the causes, frequency and stability of biotic interactions, ecologists often use field observations to generate networks of interacting species. The challenge is in determining whether unobserved interactions were undetected due to sampling, or whether they truly do not occur. This uncertainty makes it difficult to predict interactions based on ecological mechanisms. By using hierarchical Bayesian N-mixture models adapted from wildlife ecology, we differentiated the probability of detecting a species from the intensity of species interactions. We began with a brief simulation to illustrate the consequences of ignoring detectability when building species networks. We then applied our model to an empirical dataset on Andean hummingbirds to compare trait-matching and plant abundance in explaining flower visitation patterns. The daily probability of detecting a hummingbird interaction ranged from 14% to 70% among species, underlining the importance of multiple days of sampling when estimating species interactions. The trait-matching model best fit hummingbird visitation rates among plants and pollinators. The plant abundance model was inconsistent across hummingbird species and did not support neutral interactions based on increasing plant abundance. We recommend that network studies account for interaction detectability by separately modeling observation and process mechanisms, focusing on species-level discrepancy in assessing model adequacy, and estimating uncertainty when computing network statistics.

The Antarctic krill fishery is the largest in the southern ocean, but currently operates without fine-scale information on whale movement and behavior. Using a multi-year dataset of satellite-tagged whales, as well as information on krill catch levels, we analyzed the spatial distribution of whales and fisheries effort within the small-scale management units defined by the Convention for the Conservation of Antarctic Marine Living Resources (CCAMLR). Using a Bayesian movement model to partition whale movement into traveling and area-restricted search states, we found that both whale behavior and krill catch effort were spatially clustered, with distinct hotspots of the whale activity in the Gerlache and southern Branfield Straits. These areas align with increases in krill fishing effort, and present potential areas of current and future conflict. We recommend that the Antarctic West and Bransfield Strait West management units merit particular attention when setting fine-scale catch limits and, more broadly, consideration as critical areas for krill predator foraging.

Tree detection is a fundamental task in remote sensing for forestry and ecosystem ecology applications. While many individual tree segmentation algorithms have been proposed, the development and testing of these algorithms is typically site specific, with few methods evaluated against data from multiple forest types simultaneously. This makes it difficult to determine the generalization of proposed approaches, and limits tree detection at broad scales. Using data from the National Ecological Observatory Network we extend a recently developed semi-supervised deep learning algorithm to include data from a range of forest types, determine whether information from one forest can be used for tree detection in other forests, and explore the potential for building a universal tree detection algorithm. We find that the deep learning approach works well for overstory tree detection across forest conditions, outperforming conventional LIDAR-only methods in all forest types. Performance was best in open oak woodlands and worst in alpine forests. When models were fit to one forest type and used to predict another, performance generally decreased, with better performance when forests were more similar in structure. However, when models were pretrained on data from other sites and then fine-tuned using a small amount of hand-labeled data from the evaluation site, they performed similarly to local site models. Most importantly, a universal model fit to data from all sites simultaneously performed as well or better than individual models trained for each local site. This result suggests that RGB tree detection models that can be applied to a wide array of forest types at broad scales should be possible.

Functional ecology has increasingly focused on describing ecological communities based on their traits (measurable features affecting individuals fitness and performance). Analyzing trait distributions within and among forests could significantly improve understanding of community composition and ecosystem function. Historically, data on trait distributions are generated by (1) collecting a small number of leaves from a small number of trees, which suffers from limited sampling but produces information at the fundamental ecological unit (the individual); or (2) using remote sensing images to infer traits, producing information continuously across large regions, but as plots (containing multiple trees of different species) or pixels, not individuals. Remote sensing methods that identify individual trees and estimate their traits would provide the benefits of both approaches, producing continuous large-scale data linked to biological individuals. We used data from the National Ecological Observatory Network (NEON) to develop a method to scale up functional traits from 160 trees to the millions of trees within the spatial extent of two NEON sites. The pipeline consists of three stages: 1) image segmentation, to identify individual trees and estimate structural traits; 2) ensemble of models to infer leaf mass area (LMA), nitrogen, carbon, and phosphorus content using hyperspectral signatures, and DBH from allometry; and 3) predictions for segmented crowns for the full remote sensing footprint at the NEON sites.The R2 values on held out test data ranged from 0.41 to 0.75 on held out test data. The ensemble approach performed better than single partial least squares models. Carbon performed poorly compared to other traits (R2 of 0.41). The crown segmentation step contributed the most uncertainty in the pipeline, due to over-segmentation. The pipeline produced good estimates of DBH (R2 of 0.62 on held out data). Trait predictions for crowns performed significantly better than comparable predictions on pixels, resulting in improvement of R2 on test data of between to 0.26. We used the pipeline to produce individual level trait data for ∼5 million individual crowns, covering a total extent of ∼360 km2. This large dataset allows testing ecological questions on landscape scales, revealing that foliar traits are correlated with structural traits and environmental conditions.

Remote sensing can transform the speed, scale, and cost of biodiversity and forestry surveys. Data acquisition currently outpaces the ability to identify individual organisms in high resolution imagery. We outline an approach for identifying tree-crowns in RGB imagery while using a semi-supervised deep learning detection network. Individual crown delineation has been a long-standing challenge in remote sensing and available algorithms produce mixed results. We show that deep learning models can leverage existing Light Detection and Ranging (LIDAR)-based unsupervised delineation to generate trees that are used for training an initial RGB crown detection model. Despite limitations in the original unsupervised detection approach, this noisy training data may contain information from which the neural network can learn initial tree features. We then refine the initial model using a small number of higher-quality hand-annotated RGB images. We validate our proposed approach while using an open-canopy site in the National Ecological Observation Network. Our results show that a model using 434,551 self-generated trees with the addition of 2848 hand-annotated trees yields accurate predictions in natural landscapes. Using an intersection-over-union threshold of 0.5, the full model had an average tree crown recall of 0.69, with a precision of 0.61 for the visually-annotated data. The model had an average tree detection rate of 0.82 for the field collected stems. The addition of a small number of hand-annotated trees improved the performance over the initial self-supervised model. This semi-supervised deep learning approach demonstrates that remote sensing can overcome a lack of labeled training data by generating noisy data for initial training using unsupervised methods and retraining the resulting models with high quality labeled data.

Tree crown detection is a fundamental task in remote sensing for forestry and ecosystem ecology. While many individual tree segmentation algorithms have been proposed, the development and testing of these algorithms is typically site specific, with few methods evaluated against data from multiple forest types simultaneously. This makes it difficult to determine the generalization of proposed approaches, and limits tree detection at broad scales. Using data from the National Ecological Observatory Network, we extend a recently developed deep learning approach to include data from a range of forest types to determine whether information from one forest can be used for tree detection in other forests, and explore the potential for building a universal tree detection algorithm. We find that the deep learning approach works well for overstory tree detection across forest conditions. Performance was best in open oak woodlands and worst in alpine forests. When models were fit to one forest type and used to predict another, performance generally decreased, with better performance when forests were more similar in structure. However, when models were pretrained on data from other sites and then fine-tuned using a relatively small amount of hand-labeled data from the evaluation site, they performed similarly to local site models. Most importantly, a model fit to data from all sites performed as well or better than individual models trained for each local site.

A computational model describing water transport and swelling in the stratum corneum layer of the skin has been developed. The model deals with varying water diffusion coefficient within the stratum corneum layer and a moving boundary with water entering or exiting the membrane under various boundary conditions. In steady state, the model can reasonably predict the transepidermal water loss (TEWL) and the membrane thickness as a function of surface relative humidity (RH) or water activity ( a w ). The predicted TEWL is around 13 g/m 2 h in the relative humidity range of 0–70%. The model also successfully predicts nonlinear water distribution profiles within the stratum corneum at various relative humidities. In dynamic mode, the model is able to predict water distribution, membrane thickness, and water transport rate as a function of time for either hydration or dehydration under different surface boundary conditions including occlusion, exposure to air or exposure to gels. For the hydration process, the swelling rate of reaching a fully hydrated state under different surface boundary conditions is in this order: immersion in normal saline with a w =0.996>exposure to humid air with RH=0.996>dry occlusive surface. For the dehydration process, the shrinking rate from the fully hydrated state is nearly the same under the surface boundary conditions of fixed surface a w =0.05 and exposure to air with RH=5%. The dehydration rate from the fully hydrated state is faster than the reverse hydration rate for reaching the fully hydrated state.

The composition of ecological assemblages depends on a variety of factors including environmental filtering, biotic interactions and dispersal limitation. By evaluating the phylogenetic pattern of assemblages, we gain insight into the relative contribution of these mechanisms to generating observed assemblages. We address some limitations in the field of community phylogenetics by using simulations, biologically relevant null models, and cost distance analysis to evaluate simultaneous mechanisms leading to observed patterns of cooccurrence. Building from past studies of phylogenetic community structure, we applied our approach to hummingbird assemblages in the Northern Andes. We compared the relationship between relatedness and co-occurrence among predicted assemblages, based on estimates of suitable habitat and dispersal limitation, and observed assemblages. Hummingbird co-occurrence peaked at intermediate relatedness and decreased when a closely-related species was present. This result was most similar to simulations that included simultaneous effects of phylogenetic conservatism and repulsion. In addition, we found older sister taxa were only weakly more separated by geographic barriers, suggesting that time since dispersal is unlikely to be the sole factor influencing co-occurrence of closely related species. Our analysis highlights the role of multiple mechanisms acting simultaneously, and provides a hypothesis for the potential importance of competition at regional scales. COL0147267

Summary Human observation is expensive and limits the breadth of data collection. For this reason, remotely placed video cameras are increasingly used to monitor animals. One drawback of field-based video recordings is extensive review time. Computer vision can mitigate this cost and enhance data collection by extracting biological information from images with minimal time investment. MotionMeerkat is a new standalone program that identifies motion events from a video stream. After running a video, the user reviews a folder of candidate motion frames for the target organism. This tool reduces the time needed to review videos and accommodates a variety of inputs. I tested MotionMeerkat using hummingbird–plant videos recorded in a tropical montane forest. To validate the optimal parameter set for finding motion events, I counted hummingbirds observed from direct video review compared to events found in images returned from MotionMeerkat. To show the generality of the approach, MotionMeerkat was tested on a set of terrestrial and underwater videos. To assess the performance of the background subtraction for further image analysis, I hand counted the number of frames with target organisms and compared them to the MotionMeerkat output. MotionMeerkat was highly successful in finding motion events and often reduced the number of frames needed to capture hummingbird visitation by over 90%. Both background approaches effectively found a variety of organisms in ecological videos. I provide general recommendations for parameter settings and extending this approach in the future.

The Amazon basin, one of the world's core areas for biocultural diversity, includes or borders on nine South American states. The remote and biodiverse Amazon borderlands shared by these states contain over 12,000 kilometers of international boundaries and are increasingly threatened by transboundary infrastructure initiatives. This paper combines geographic information systems (GIS), field observations, and document research to investigate the relationship between cultural diversity and the Amazon borderlands: (1) Are the borderlands more culturally diverse than the Amazonian countries and Amazonian lowland rainforest biome? (2) If so, what characterizes this diversity? Results introduce the unique characteristics of the Amazon borderlands and underscore the argument for an alternative means of Amazon integration based on standing forest and biocultural diversity.

Remote sensing can transform the speed, scale, and cost of biodiversity and forestry surveys. Data acquisition currently outpaces the ability to identify individual organisms in high resolution imagery. We outline an approach for identifying tree-crowns in RGB imagery while using a semi-supervised deep learning detection network. Individual crown delineation has been a long-standing challenge in remote sensing and available algorithms produce mixed results. We show that deep learning models can leverage existing Light Detection and Ranging (LIDAR)-based unsupervised delineation to generate trees that are used for training an initial RGB crown detection model. Despite limitations in the original unsupervised detection approach, this noisy training data may contain information from which the neural network can learn initial tree features. We then refine the initial model using a small number of higher-quality hand-annotated RGB images. We validate our proposed approach while using an open-canopy site in the National Ecological Observation Network. Our results show that a model using 434,551 self-generated trees with the addition of 2848 hand-annotated trees yields accurate predictions in natural landscapes. Using an intersection-over-union threshold of 0.5, the full model had an average tree crown recall of 0.69, with a precision of 0.61 for the visually-annotated data. The model had an average tree detection rate of 0.82 for the field collected stems. The addition of a small number of hand-annotated trees improved the performance over the initial self-supervised model. This semi-supervised deep learning approach demonstrates that remote sensing can overcome a lack of labeled training data by generating noisy data for initial training using unsupervised methods and retraining the resulting models with high quality labeled data.

By specialising on specific resources, species evolve advantageous morphologies to increase the efficiency of nutrient acquisition. However, many specialists face variation in resource availability and composition. Whether specialists respond to these changes depends on the composition of the resource pulses, the cost of foraging on poorly matched resources, and the strength of interspecific competition. We studied hummingbird bill and plant corolla matching during seasonal variation in flower availability and morphology. Using a hierarchical Bayesian model, we accounted for the detectability and spatial overlap of hummingbird-plant interactions. We found that despite seasonal pulses of flowers with short-corollas, hummingbirds consistently foraged on well-matched flowers, leading to low niche overlap. This behaviour suggests that the costs of searching for rare and more specialised resources are lower than the benefit of switching to super-abundant resources. Our results highlight the trade-off between foraging efficiency and interspecific competition, and underline niche partitioning in maintaining tropical diversity.

Identifying which species are at greatest risk, what makes them vulnerable, and where they are distributed are central goals for conservation science. While knowledge of which factors influence extinction risk is increasingly available for some taxonomic groups, a deeper understanding of extinction correlates and the geography of risk remains lacking. Here, we develop a predictive random forest model using both geospatial and mammalian species' trait data to uncover the statistical and geographic distributions of extinction correlates. We also explore how this geography of risk may change under a rapidly warming climate. We found distinctive macroecological relationships between species-level risk and extinction correlates, including the intrinsic biological traits of geographic range size, body size and taxonomy, and extrinsic geographic settings such as seasonality, habitat type, land use and human population density. Each extinction correlate exhibited ranges of values that were especially associated with risk, and the importance of different risk factors was not geographically uniform across the globe. We also found that about 10% of mammals not currently recognized as at-risk have biological traits and occur in environments that predispose them towards extinction. Southeast Asia had the most actually and potentially threatened species, underscoring the urgent need for conservation in this region. Additionally, nearly 40% of currently threatened species were predicted to experience rapid climate change at 0.5 km/year or more. Biological and environmental correlates of mammalian extinction risk exhibit distinct statistical and geographic distributions. These results provide insight into species-level patterns and processes underlying geographic variation in extinction risk. They also offer guidance for future conservation research focused on specific geographic regions, or evaluating the degree to which species-level patterns mirror spatial variation in the pressures faced by populations within the ranges of individual species. The added impacts from climate change may increase the susceptibility of at-risk species to extinction and expand the regions where mammals are most vulnerable globally.

Comparison of the taxonomic, phylogenetic, and trait dimensions of beta diversity may uncover the mechanisms that generate and maintain biodiversity, such as geographic isolation, environmental filtering, and convergent adaptation. We developed an approach to predict the relationship between environmental and geographic distance and the dimensions of beta diversity. We tested these predictions using hummingbird assemblages in the northern Andes. We expected taxonomic beta diversity to result from recent geographic barriers limiting dispersal, and we found that cost distance, which includes barriers, was a better predictor than Euclidean distance. We expected phylogenetic beta diversity to result from historical connectivity and found that differences in elevation were the best predictors of phylogenetic beta diversity. We expected high trait beta diversity to result from local adaptation to differing environments and found that differences in elevation were correlated with trait beta diversity. When combining beta diversity dimensions, we observe that high beta diversity in all dimensions results from adaption to different environments between isolated assemblages. Comparisons with high taxonomic, low phylogenetic, and low trait beta diversity occurred among lowland assemblages separated by the Andes, suggesting that geographic barriers have recently isolated lineages in similar environments. We provide insight into mechanisms governing hummingbird biodiversity patterns and provide a framework that is broadly applicable to other taxonomic groups.

Premise of the study: Low-elevation surveys with small aerial drones (micro–unmanned aerial vehicles [UAVs]) may be used for a wide variety of applications in plant ecology, including mapping vegetation over small- to medium-sized regions. We provide an overview of methods and procedures for conducting surveys and illustrate some of these applications. Methods: Aerial images were obtained by flying a small drone along transects over the area of interest. Images were used to create a composite image (orthomosaic) and a digital surface model (DSM). Vegetation classification was conducted manually and using an automated routine. Coverage of an individual species was estimated from aerial images. Results: We created a vegetation map for the entire region from the orthomosaic and DSM, and mapped the density of one species. Comparison of our manual and automated habitat classification confirmed that our mapping methods were accurate. A species with high contrast to the background matrix allowed adequate estimat...

The taxonomic, phylogenetic and trait dimensions of beta diversity each provide us unique insights into the importance of historical isolation and environmental conditions in shaping global diversity. These three dimensions should, in general, be positively correlated. However, if similar environmental conditions filter species with similar trait values, then assemblages located in similar environmental conditions, but separated by large dispersal barriers, may show high taxonomic, high phylogenetic, but low trait beta diversity. Conversely, we expect lower phylogenetic diversity, but higher trait biodiversity among assemblages that are connected but are in differing environmental conditions. We calculated all pairwise comparisons of approximately 110 × 110 km grid cells across the globe for more than 5000 mammal species (approx. 70 million comparisons). We considered realms as units representing geographical distance and historical isolation and biomes as units with similar environmental conditions. While beta diversity dimensions were generally correlated, we highlight geographical regions of decoupling among beta diversity dimensions. Our analysis shows that assemblages from tropical forests in different realms had low trait dissimilarity while phylogenetic beta diversity was significantly higher than expected, suggesting potential convergent evolution. Low trait beta diversity was surprisingly not found between isolated deserts, despite harsh environmental conditions. Overall, our results provide evidence for parallel assemblage structure of mammal assemblages driven by environmental conditions at a global scale.

Evaluation of the performance of a paper machine control system was carried out using statistical methods. A sequential approach to the evaluation and optimization of control systems was used to evaluate and optimize control system performance. Time series techniques were used to determine the potential for improved process control. Spectral analysis techniques and simple open-loop process identification experiments were used to find the cause of poor control system performance. Significant improvements in control system performance were realized.