Developing a modern data workflow for regularly updated data.

Over the past decade, biology has undergone a data revolution in how researchers collect data and the amount of data being collected. An emerging challenge that has received limited attention in biology is managing, working with, and providing access to data under continual active collection. Regularly updated data present unique challenges in quality assurance and control, data publication, archiving, and reproducibility. We developed a workflow for a long-term ecological study that addresses many of the challenges associated with managing this type of data. We do this by leveraging existing tools to 1) perform quality assurance and control; 2) import, restructure, version, and archive data; 3) rapidly publish new data in ways that ensure appropriate credit to all contributors; and 4) automate most steps in the data pipeline to reduce the time and effort required by researchers. The workflow leverages tools from software development, including version control and continuous integration, to create a modern data management system that automates the pipeline.

A general deep learning model for bird detection in high-resolution airborne imagery.

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.

Long-term monitoring and experimental manipulation of a Chihuahuan desert ecosystem near Portal, Arizona (1977-2013).

Desert ecosystems have long served as model systems in the study of ecological concepts (e.g., competition, resource pulses, top-down/bottom-up dynamics). However, the inherent variability of resource availability in deserts, and hence consumer dynamics, can also make them challenging ecosystems to understand. Study of a Chihuahuan desert ecosystem near Portal, Arizona began in 1977. At this site, 24 experimental plots were established and divided among controls and experimental manipulations. Experimental manipulations over the years include removal of all or some rodent species, all or some ants, seed additions, and various alterations of the annual plant community. This dataset includes data previously available through an older data publication and adds 11 years of data. It also includes additional ant and weather data not previously available. These data have been used in a variety of publications documenting the effects of the experimental manipulations as well as the response of populations and communities to long-term changes in climate and habitat. Sampling is ongoing and additional data will be published in the future.

Shifts in competitive structures can drive variation in species' phenology.

For many species, a well documented response to anthropogenic climate change is a shift in various aspects of its life history, including its timing or phenology. Often, these phenological shifts are associated with changes in abiotic factors used as proxies for resource availability or other suitable conditions. Resource availability, however, can also be impacted by competition, but the impact of competition on phenology is less studied than abiotic drivers. We fit generalized additive models (GAMs) to a long-term experimental dataset on small mammals monitored in the southwestern United States and show that altered competitive landscapes can drive shifts in breeding timing and prevalence, and that, relative to a dominant competitor, other species exhibit less specific responses to environmental factors. These results suggest that plasticity of phenological responses, which is often described in the context of annual variation in abiotic factors, can occur in response to biotic context as well. Variation in phenological responses under different biotic conditions shown here further demonstrates that a more nuanced understanding of shifting biotic interactions is useful to better understand and predict biodiversity patterns in a changing world.

Strong self-limitation promotes the persistence of rare species.

Theory has recognized a combination of niche and neutral processes each contributing, with varying importance, to species coexistence. However, long-term persistence of rare species has been difficult to produce in trait-based models of coexistence that incorporate stochastic dynamics, raising questions about how rare species persist despite such variability. Following recent evidence that rare species may experience significantly different population dynamics than dominant species, we use a plant community model to simulate the effect of disproportionately strong negative frequency dependence on the long-term persistence of the rare species in a simulated community. This strong self-limitation produces long persistence times for the rare competitors, which otherwise succumb quickly to stochastic extinction. The results suggest that the mechanism causing species to be rare in this case is the same mechanism allowing those species to persist.