Vermont Project: Adaptive Management with AMMonitor

Vermont Project

Adaptive Management with AMMonitor

August 2020 - August 2026

Personnel

Participating Agencies

USGS

Adaptive management (AM) is a cyclical process that improves natural resource management and generates knowledge that can inform future decisions. Autonomous monitoring units (AMUs) such as camera traps or recording devices provide opportunities to collect large volumes of species’ occurrence data over long periods of time with relatively low cost. However, the data management requirements of an AMU-based monitoring effort can be immense: processing audio/image data to generate detections can be complicated, and integrating streams of AMU data into species’ distribution models or other analyses may appear unattainable.

To operationalize AM, the USGS Vermont Cooperative Fish and Wildlife Research Unit created the open-source R package AMMonitor, which simplifies the process of moving from remotely collected data to species’ status, trend analysis, and reports.

The objectives of this project are to 1) Strengthen and expand the AMMonitor community, and 2) Develop the photographic monitoring capacity of AMMonitor. The Vermont Cooperative Fish and Wildlife Research Unit is working with USGS ScienceBase to facilitate and document monitoring files as a permanent data repository. Further, we are working with USGS Cloud Hosting to develop CNN models to identify species within monitoring files (images, audio) in an automated fashion using machine learning algorithms. The results of this work should streamline data analysis and ultimately permit adaptive management of natural resources in an expedient manner.

Research Publications

Research Publications	Publication Date
Donovan. T. M., C. Balantic, J. Katz, M. Massar, R. Knutson, K. Duh, P. W. Jones, K. Epstein, J. Lacasse-Roger, and J. Dias. 2021. Remote Ecological Monitoring with Smartphones and Tasker. Journal of Fish and Wildlife Management 12:163–173. \| Abstract \| Download \| Publisher Website	June 2021	Researchers have increasingly used autonomous monitoring units to record animal sounds, track phenology with timed photographs, and snap images when triggered by motion. We piloted the use of smartphones to monitor wildlife in the Riverside East Solar Energy Zone (California) and at Indiana Dunes National Park (Indiana). For both efforts, we established remote autonomous monitoring stations in which we housed an Android smartphone in a weather-proof box mounted to a pole and powered by solar panels. We connected each smartphone to a Google account, and the smartphone received its recording/photo schedule daily via a Google Calendar connection when in data transmission mode. Phones were automated by Tasker, an Android application for automating cell phone tasks. We describe a simple approach that could be adopted by others who wish to use nonproprietary methods of data collection and analysis.
Clarfeld, L., A. Sirén, B. Mulhall, T. Wilson, E. Bernier, J. Farrell, G. Lunde, N. Hardy, K. Gieder, R. Abrams, S. Staats, S. McLellan, and T. Donovan. 2023. Evaluating a tandem human-machine approach to labelling of wildlife in remote camera monitoring. Ecological Informatics 77:e102257 https://doi.org/10.1016/j.ecoinf.2023.102257. \| Abstract \| Download \| Publisher Website	August 2023	Remote cameras (“trail cameras”) are a popular tool for non-invasive, continuous wildlife monitoring, and as they become more prevalent in wildlife research, machine learning (ML) is increasingly used to automate or accelerate the labor-intensive process of labelling (i.e., tagging) photos. Human-machine hybrid tagging approaches have been shown to greatly increase tagging efficiency (i.e., time to tag a single image). However, those potential increases hinge on the extent to which an ML model makes correct vs. incorrect predictions. We performed an experiment using a ML model that produces bounding boxes around animals, people, and vehicles in remote camera imagery (MegaDetector) to consider the impact of a ML model's performance on its ability to accelerate human labeling. Six participants tagged trail camera images collected from 12 sites in Vermont and Maine, USA (January–September 2022) using three tagging methods (one with ML bounding box assistance and two without assistance). We used a generalized linear mixed model to examine the influence of ML model performance and tagging method on tagging efficiency. We found that ML bounding boxes offer significant improvement in tagging efficiency when labelling data compared to unassisted tagging. Additionally, the time taken to label with bounding boxes was not statistically different from an unassisted tagging approach. However, we found that gains in efficiency are contingent on the ML algorithm's performance and that incorrect ML predictions, particularly the 4.2% false positive and 3.6% false negative predictions, can slow the tagging process compared to a non-hybrid approach. These findings indicate that although practitioners usually forgo the production of bounding boxes when selecting a data labelling process due to the increased effort, ML bounding box-assisted tagging can offer an efficient method for labeling. More broadly, ML-assisted data labelling offers an opportunity to accelerate the analysis of trail camera imagery, but an assessment of the ML model's performance can illuminate whether the hybrid-tagging approach is ultimately a help or hinderance.
Balantic, C., and T. M. Donovan. 2019. Temporally-adaptive acoustic sampling to maximize detection across a suite of focal wildlife species. Ecology and Evolution 9(18)10582-10600. DOI: 10.1002/ece3.5579 \| Abstract \| Download \| Publisher Website	August 2019	1. Acoustic recordings of the environment can produce species presence–absence data for characterizing populations of sound‐producing wildlife over multiple spatial scales. If a species is present at a site but does not vocalize during a scheduled audio recording survey, researchers may incorrectly conclude that the species is absent (“false negative”). The risk of false negatives is compounded when audio devices have sampling constraints, do not record continuously, and must be manually scheduled to operate at pre‐selected times of day, particularly when research programs target multiple species with acoustic availability that varies across temporal conditions.<br><br>2. We developed a temporally adaptive acoustic sampling algorithm to maximize detection probabilities for a suite of focal species amid sampling constraints. The algorithm combines user‐supplied species vocalization models with site‐specific weather forecasts to set an optimized sampling schedule for the following day. To test our algorithm, we simulated hourly vocalization probabilities for a suite of focal species in a hypothetical monitoring area for the year 2016. We conducted a factorial experiment that sampled from the 2016 acoustic environment to com ‐pare the probability of acoustic detection by a fixed (stationary) schedule versus a temporally adaptive optimized schedule under several sampling efforts and monitoring durations.<br><br>3. We found that over the course of a study season, the probability of acoustically capturing a focal species (given presence) at least once via automated acoustic monitoring was greater (and acoustic capture occurred earlier in the season) when using the temporally adaptive optimized schedule as compared to a fixed schedule.<br><br>4. The advantages of a temporally adaptive optimized acoustic sampling schedule are magnified when a study duration is short, sampling effort is low, and/or species acoustic availability is minimal. This methodology presents the opportunity to maximize acoustic monitoring sampling efforts amid constraints.
Balantic, C., and T. M. Donovan. 2019. Statistical learning mitigation of false positive detections in automated acoustic wildlife monitoring. Bioacoustics 29(3):296-321. DOI: 10.1080/09524622.2019.1605309 \| Abstract \| Download \| Publisher Website	May 2019	Audio sampling of the environment can provide long-term, landscape-scale presence-absence data to model populations of sound-producing wildlife. Automated detection systems allow researchers to avoid manually searching through large volumes of recordings, but often produce unacceptable false positive rates. We developed methods that allow researchers to improve template-based automated detection using a suite of statistical learning algorithms when false positive rates are problematic. To test our method, we acquired 668 hours of recordings in the Sonoran Desert, California USA between March 2016 and May 2017, and created spectrogram cross-correlation templates for three target avian species. We trained and tested five classification algorithms and four performance-weighted ensemble classifier methods on target signals and false alarms from March 2016, and then selected high-performing ensemble classifiers from the train/test phase to predict the class of new detections thereafter. For three target species, our ensemble classifiers were able to identify 98%, 81%, and 100% of false alarms compared with the baseline template detection system, and comparative positive predictive values improved from 6% to 69%, 87% to 95%, and 2% to 77%. We show that statistical learning approaches can be implemented to mitigate false detections acquired via template-based automated detection in automated acoustic wildlife monitoring.
Balantic, C., and T. M. Donovan. 2019. Dynamic wildlife occupancy models using automated acoustic monitoring data. Ecological Applications 29(3):e01854. \| Abstract \| Download \| Publisher Website	April 2019	Automated acoustic monitoring of wildlife has been used to characterize populations of sound‐producing species across large spatial scales. However, false negatives and false positives produced by automated detection systems can compromise the utility of these data for researchers and land managers, particularly for research programs endeavoring to describe colonization and extinction dynamics that inform land use decision‐making. To investigate the suitability of automated acoustic monitoring for dynamic occurrence models, we simulated underlying occurrence dynamics, calling patterns, and the automated acoustic detection process for a hypothetical species under a range of scenarios. We investigated an automated species detection aggregation method that considered a suite of options for creating encounter histories. From these encounter histories, we generated parameter estimates and computed bias for occurrence, colonization, and extinction rates using a dynamic occupancy modeling framework that accounts for false positives via small amounts of manual confirmation. We were able to achieve relatively unbiased estimates for all three state parameters under all scenarios, even when the automated detection system was simulated to be poor, given particular encounter history aggregation choices. However, some encounter history aggregation choices resulted in unreliable estimates; we provide caveats for avoiding these scenarios. Given specific choices during the detection aggregation process, automated acoustic monitoring data may provide an effective means for tracking species occurrence, colonization, and extinction patterns through time, with the potential to inform adaptive management at multiple spatial scales.
Balantic, C., and T. Donovan. 2020. AMMonitor: Remote monitoring of biodiversity in an adaptive framework with R. Methods in Ecology and Evolution 11:869-877. DOI: 10.1111/2041-210X.13397 \| Abstract \| Download \| Publisher Website	April 2020	1. Ecological research and management programs are increasingly using autonomous monitoring units (AMUs) to collect large volumes of acoustic and/or photo data to address pressing management objectives or research goals. The data management requirements of an AMU-based monitoring effort are often overwhelming, with a considerable amount of processing to translate raw data into models and analyses that have research and management utility.<br><br>2. We created the r package AMMonitor to simplify the process of moving from remotely collected data to analysis and results, using a comprehensive SQLite database for data management that tracks all components of a remote monitoring program. This framework enables the tracking of analyses and research/ management objectives through time.<br><br>3. We illustrate the AMMonitor approach with the example of evaluating an occurrence-based management objective for a target species. First, we provide an overview of the database and data management approach. Next, we illustrate a few available workflows: temporally adaptive sampling, automated detection of species sounds from acoustic recordings and aggregation of automated detections into an encounter history for use in an occupancy analysis, the outcome of which can be analysed with respect to the motivating management objective.<br><br>4. Without a comprehensive framework for efficiently moving from raw remote monitoring data collection to results and analysis, monitoring programs are limited in their capacity to systematically characterize ecological processes and inform management decisions through time. AMMonitor provides an option for such a framework. Code, comprehensive documentation and step-by-step examples are available online at https://code.usgs.gov/vtcfwru/AMMonitor

Data or Software Releases

Type	Citation	Publication Date
Data Release	Wilson, T. L., Siren, A. P. K., Berube, J. A., Morrow, C. M., Wattles, D. W., Huguenin, M., Clarfeld, L. A., Huber, K. E., Donovan, T. M., 2024, Massachusetts Wildlife Monitoring Project (2022 - 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P13UNTFB.	June 2024
Data Release	Webb, S.M., Clarfeld, L.A., Huber, K.E., and Donovan. T.M., 2024, Maine Department of Inland Fisheries and Wildlife Volume 1 (2022 - 2023): U.S. Geological Survey data release, https://doi.org/10.5066/P9ML66K3.	April 2024
Data Release	Prout, L. S., Siren, A. P. K., Callahan, C. B., Wilson, T. L., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, USDA White Mountain National Forest Volume 1 (2014 - 2024): U.S. Geological Survey data release, doi:10.5066/P1PUEYQK.	June 2024
Data Release	Patry, R. K., Sirén, A. P. K., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Dartmouth College Woodlands Wildlife Monitoring Project Volume 1 (2014 - 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P133GQPK.	May 2024
Data Release	Morelli, T. L., Sirén, A. P. K., Patry, R. K., Cliché, R. M., Clark, J. P., Loesberg, J. A., Courtot, K. A., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, SiMPL Wildlife Magnet Project Data Release Volume 1 (2016 - 2023): U.S. Geological Survey data release, doi:10.5066/P1VUKJQK.	May 2024
Data Release	Kantar, L. E., Sirén, A. P. K., Wilson, T. L., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Maine Department of Inland Fisheries and Wildlife Moose Project - Volume 2 (2021 - 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P132SU4S.	June 2024
Data Release	Jones, H., Sirén, A. P. K., Callahan, C. B., Holman, H., Marchand, M. N., Kilborn, J. R., Wilson, T. L., Morelli, T. L., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, New Hampshire Fish and Game Department Volume 1 (2014 - 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P13ISNAI.	September 2024
Data Release	Jensen, P. G., Rojas, V. G., Frair, J. L., McNulty, S. A., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Adirondack Inventory & Monitoring (AIM) Network Data Release Volume 1 (2021 - 2023): U.S. Geological Survey data release, doi:10.5066/P1K6FWJ4.	May 2024
Data Release	Gieder, K. D., Bernier, C. A., Staats, S. A., Wixsom, S. J., Abrams, R. J., Cahill, J. R., Crumley, K., Royar, K., Siren, A. P. K., Wilson, T. L., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, USDA Green Mountain National Forest Volume 1 (2016 - 2022): U.S. Geological Survey data release, doi:10.5066/P1GVIBFL.	June 2024
Data Release	Gieder, K. D., Bernier, C. A., Royar, K., Siren, A. P. K., Crumley, K., Smith, T. R., Courtot, K., Wilson, T. L., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Vermont Fish and Wildlife Department Volume 1 (2014 - 2022): U.S. Geological Survey data release, https://doi.org/10.5066/P14MFBJT.	June 2024
Data Release	Cliche, R. M., Sirén, A. P. K., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Silvio O Conte National Fish and Wildlife Refuge Wildlife Monitoring Project (2014 - 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P18M87XP.	May 2024
Data Release	Clarfeld, L., Donovan, T., Siren, A., Mulhall, B., Bernier, E., Farrell, J., Lunde, G., Hardy, N., Abrams, R., Staats, S., and McLellan, S., 2023, Evaluating a tandem human-machine approach to labelling of wildlife in remote camera monitoring: U.S. Geological Survey data release, https://doi.org/10.5066/P9FGUQEZ.	August 2023
Software Release	Clarfeld, L., C. Tang, K. Huber, C. Balantic, and T. Donovan. AMMonitor: Remote monitoring of biodiversity in an adaptive framework. Version 2.1: U.S. Geological Survey software release. Reston, VA. https://doi.org/10.5066/P13MRDRV	October 2024
Software Release	Clarfeld, L., C. Tang, K. Huber, C. Balantic, and T. Donovan. AMMonitor: Remote monitoring of biodiversity in an adaptive framework. Version 2.0.0: U.S. Geological Survey software release. Reston, VA. https://doi.org/10.5066/P1DKQ5RP	June 2024
Data Release	Balantic, C. M., Knutson, R., Duh, K., Clarfeld, L. A., Huber, K. E., and Donovan, T. M., 2024, Indiana Dunes National Park Volume 1 (2019): U.S. Geological Survey data release, doi:10.5066/P1CXDATH.	June 2024

Cooperative Fish and Wildlife Research Units Program: Vermont Education, Research and Technical Assistance for Managing Our Natural Resources

Vermont Project

Adaptive Management with AMMonitor

August 2020 - August 2026

Personnel

Participating Agencies

Cooperative Fish and Wildlife Research Units Program: Vermont
Education, Research and Technical Assistance for Managing Our Natural Resources