- acoustic monitoring
- acoustic analysis
- capture and handling training
- acoustics training
Monitoring: Do you need an area assessed for bat activity?
Dr. Cori Lausen [left] and colleagues from the University of Calgary [including Erin Baerwald, right] produced the first North American protocol for bat activity assessment in the pre-construction phase of wind farm development.
Newer, taller wind turbines seem to be killing more bats than the older style turbines [Barclay et al. 2007]. Strategic placement of new wind farms may reduce bat mortality. Pre-construction monitoring for bats can help determine whether a proposed wind farm location is in the migratory path of bats. After turbines are built, post-construction monitoring is necessary to determine bat mortality rates. Mitigation strategies include shutting turbines down in low wind speeds.
Acoustic techniques are still the most effective tool for pre-construction monitoring; while radar can be useful for counting objects flying in a wind development area, bats and birds are not easily distinguished. While habitat features can be useful for determining potential hibernation and maternity sites, bat migration routes are difficult to predict. The wind farms pictured here in SW Alberta have higher bat mortality rates than turbines further east in that province. Crop fields don’t look like typical bat habitat, which makes pre-construction acoustic survey critical.
Acoustic Analysis: Do you need acoustic files analyzed?
Analysis of bat acoustic files is an acquired skill. Bats are not birds. In other words, their calls are not designed to be species specific to attract mates; instead they are functional, designed to catch insect prey. Bats of similar body size that feed on similar sizes of insects can have very similar echolocation calls. And there is a great deal of plasticity in echolocation calls, allowing bats to change their echolocation call structure to best suit their surroundings. For example, near trees, an individual bat will produce a different call structure than when it is foraging in an open field. As such, only after examining many recordings of bats flying in varied environments, can one most confidently differentiate species. And even then, some species of bats are very difficult and sometimes impossible to tell apart 100% of the time. As automated call analysis becomes available, it will be important that results be interpreted with an appropriate level of skepticism, such that species identifications and probabilities are not blindly accepted.
Dr. Cori Lausen offers training sessions in how to analyze your own calls, but encourages all students to acquire experience following training.
Survey: Do you need a bat survey conducted and/or training on how to conduct your own surveys?
Capture and detection probabilities are different among species, necessitating the use of both survey techniques. How and where mistnets are deployed can severely bias species capture ratios. For example, in the Pacific Northwest, mistnetting over standing water will result in a high capture bias of little brown bats, while netting in cluttered forest trails will bias captures to long-eared bats and California Myotis. Nets low to the ground have less of a chance of catching high-flying species such as big browns, hoary bats, red bats and spotted bats.
As White Nose Syndrome continues to spread in North America, it is becoming evident that there is a limited amount of time in which to conduct surveys in western North America. Many areas in western North America have yet to be surveyed to determine bat biodiversity.
Wish to do your own capture surveys? I can provide training in bat capture and handling. Pre-exposure rabies vaccinations are required. Photo above shows Cori measuring a bat in Nahanni National Park, NWT. In 2006, she conducted the first large scale bat survey in the Northwest Territories, increasing known species from 3 to 7.
Bats play an important role in ecosystems; they are the only major consumer of night-time insects; they are critical pollinators in some areas, seed-dispersers in others; and more.
There are vast regions around the globe that remain unsurveyed for bat biodiversity. Despite the fact that bats comprise more than one quarter of all mammal species [there are >1000 species of bats world-wide], much remains unknown about these naturally cryptic creatures. Baseline species diversity information for many areas of North America is lacking, and is likely to change dramatically with climate change; documenting biodiversity and following changes in species ranges requires formal surveys.
Training: Does your company need training on acoustics techniques and/or analysis?
There is a growing demand for knowing how to conduct acoustic monitoring of bats. If you are new to the field of bat acoustics, the choices of detectors and software can seem overwhelming. Each detector offers advantages over others depending on the goals of your monitoring. How you plan to analyze your data also depends on your end goals and will dictate what types of files you want to generate from your detector and what software package would be most suitable for your situation.
There are basically 2 types of bat detectors suitable for monitoring passively at wind development areas: full spectrum and zero-crossing. The advantage of using full spectrum is that it digitizes the ultrasound by sampling at high rates (>190 kHz), retaining all aspects of the sound including multiple harmonics and amplitude. The main disadvantage of recording full spectrum is that sound files are very large (1-2 megabytes per bat pass). Most full spectrum detectors provide an option for compressing the wave files during recording, but when uncompressed during downloading, memory demand remains high. File conversion times can also be substantial.
Zero-crossing bat detection, specifically the Anabat detector, works by counting each time a sound wave passes the zero point (imaginary line drawn through the middle of a sound wave); at a preset number of crossings (called Division Ratio, often 8 or 16), a time measurement is made (allowing frequency to be recorded). This provides a time-frequency output of the ultrasound, recording only one frequency, the most intense one. As such, only one harmonic is displayed at any one time, and no amplitude data are retained. The main advantage of this system over full spectrum is that file sizes are significantly smaller (2-5 kilobytes per bat pass), requiring small memory cards in detectors, much shorter download time in the field, and less requirement for back-up hard-drives; additionally, zero-crossing detectors use less energy during operation than full spectrum units. However, the primary reason to use full spectrum over zero-crossing in some situations is that full spectrum may allow for better species differentiation of acoustically similar species. Unfortunately, there are currently no published studies comparing these two types of detectors’ abilities to resolve species, and until this happens, it is difficult to conclude how much detail about of the original ultrasound needs to be recorded to sufficiently meet various monitoring goals. If full spectrum does not increase species identification to a large degree in your area, you may wish to use zero-crossing instead and spend minutes instead of hours downloading data while in the field.
As passive monitoring of wind development areas continues to generate extremely large datasets of files, automated identification of files (noise vs. bat and file labels indicating bat species or species groups) is being sought. Two popular software packages currently offer auto-ID options: AnalookW (for zero-crossing data) and Sonobat (for full spectrum data).
Only one bat detector, the SM2Bat, produces files that can be analyzed in either software package. Auto-ID is in its infancy and does not come without a set of inherent problems: bats use ultrasound as a functional tool to navigate and find insect prey, and as such, similar sized bats in similar environmental situations (degree of clutter) can produce similar echolocation calls, making differentiation difficult. However, used cautiously, auto-ID software can provide bat biologists with substantial time-savings by not having to view each file. Sonobat uses discriminant function analysis (DFA) and other hierarchical decisions to arrive at labels for files, while AnalookW uses filters that match pulses in files to a set of criteria that describe the shape and frequency of bat pulses. Other auto-ID software packages are on the horizon (e.g. BCID [Bat Call Identification Inc.], SongScope [Wildlife Acoustics]).
Fundamental to conservation of biodiversity is knowing what species exist in an area!