We have made available our first set of results for one of three status variables for North American bats. Through this work we developed an analytical pipeline supported by web-based infrastructure for integrating continental scale bat monitoring data (stationary acoustic, mobile acoustic, and capture records) to estimate summer (May 1–Aug 31) occupancy probabilities and changes in occupancy over time for 12 North American bat species. This serves as one of multiple lines of evidence that inform the status and trends of bat populations. We analyzed data from a total of 12 bat species, 11 of which have tested positive for Pseudogymnoascus destructans (Pd), a fungal pathogen that causes white-nose syndrome (WNS)—a disease that has led to significant rates of mortality for subterranean hibernating bat species in North America. A twelfth species was also selected because of high rates of mortality at wind energy facilities. Additional species were considered but not selected due to data limitations.
We estimated occupancy probabilities for 2010 through 2019 for three species (Myotis lucifugus, MYLU; Myotis septentrionalis, MYSE; and Perimyotis subflavus, PESU). For an additional nine species, we estimated occupancy probabilities for 2016 through 2019 (Myotis evotis, MYEV; Myotis grisescens, MYGR; Myotis leibii, MYLE; Myotis thysanodes, MYTH; Myotis volans, MYVO; Myotis yumanensis, MYYU; Eptesicus fuscus, EPFU; Lasionycteris noctivagans, LANO; and Lasiurus cinereus, LACI). For each species, we provide range-wide occupancy probability predictions (e.g., predicted summer occupancy distribution maps) each year and provide regional estimates of mean occupancy probability aggregated at a variety of spatial scales (state/province/territory, range-wide). For each species, we also provide trends over time (average annual change rate and total change rate) in mean occupancy probabilities at multiple spatial scales (state/province/territory, range-wide) and when possible, over multiple timescales. The representativeness of sampling data for each species’ status and trend estimates (e.g., state/province/territory) were also evaluated based on the percent of grid cells sampled each year with a goal of understanding the reliability of regional estimates and improving future monitoring efforts.
Results suggest that over the short-term (2016-2019), two (Myotis lucifugus and Perimyotis subflavus) of 12 species have experienced declines in range-wide average occupancy probability with at least 95% certainty. Seven species showed either minor increases or decreases in range-wide average occupancy probability but with less than 95% certainty in both trend indicators. Results over the longer term (eight years and 10 years of sampling) suggest that three hibernating species known to be highly affected by white-nose syndrome (Myotis lucifugus, Myotis septentrionalis, and Perimyotis subflavus) have experienced marked declines in range-wide average occupancy probabilities, with severity varying by species and region. Finally, the results for three species (Eptesicus fuscus, Lasiurus cinereus, Lasionycteris noctivagans) were inconclusive due to 1) borderline convergence issues in the model fitting procedure which suggests potentially unreliable estimates, 2) failure to reliably distinguish between false positives and true positive detections for ambiguous detections, and 3) largely uninformative covariates for occupancy and detection. For Myotis lucifugus, Myotis septentrionalis, and Perimyotis subflavus we found meaningful associations in space and time between declining winter populations (likely a result of WNS) and summer occupancy distributions.
This work represents the most comprehensive effort to date to model North American bat distributions across their continental ranges. Despite current limitations highlighted in the discussion of the report (see page 86 through 89), the analytical methods and resulting status and trends estimates provide the best available science on summer bat populations across North America and will continue to improve over time as monitoring data sets and analytical methods improve. Moving forward, our occupancy analyses will continue to improve with submission of more 1) data from currently underrepresented areas (i.e., improved geographic representation), 2) manually-vetted acoustic recordings, 3) capture records, and 4) roost location and count data (summer and winter).
You can access results through the products listed below. Please note that information accessed through these products do not satisfy agency consultation requirements under Section 7(a)(2) of the U.S. Endangered Species Act of 1973 (87 Stat. 884, as amended 16 U.S.C. 1531 et seq.; ESA). For consultation purposes in the U.S., use the U.S. Fish and Wildlife Service's IPaC program, or contact your local field office.
An online summary hosted on the NABat Partner Portal where you will find high resolution occupancy and change maps along with trend estimates for how occupancy has changed over multiple time periods (2016-2019, 2012-2019, and 2010-2019) across the modeled species range and for individual states and provinces.
A USGS Data Release where you can access the .csv’s of the model results that were used to generate these maps and trend estimates.
Interactive maps supplied through a Tableau dashboard where you can manipulate your view and toggle between years and species.
A detailed report that includes
Acknowledgements
Executive Summary
1. Introduction and Purpose
2. Methods - Datasets
2.1 Response data
2.2 Covariates
2.3 Range maps
3. Results
3.1 Useful definitions for interpreting results
3.2 Organization of results
3.3 Myotis lucifugus
3.4 Myotis septentrionalis
3.5 Perimyotis subflavus
3.6 Myotis evotis
3.7 Myotis grisescens
3.8 Myotis leibii
3.9 Myotis thysanodes
3.10 Myotis volans
3.11 Myotis yumanensis
3.12 Inconclusive Results
3.12.1 Eptesicus fuscus
3.12.2 Lasionycteris noctivagans
3.12.3 Lasiurus cinereus
4. Discussion
4.1 Interpreting occupancy estimates and trends
4.2 Data limitations
4.3 Analytical limitations
4.4 Looking forward
References
Appendix A: Statistical Method Used for the False-positive Occupancy Modeling and Predictions
Appendix B: State/Province/Territory Level Results
Appendix C: Occupancy Modeling Covariate Effects
Appendix D: Comparing predicted occupancy probabilities to monitoring data
If you have feedback to offer on these products that can help improve future iterations, let us know.
Comentarios