Wind power, the transformation of energy from moving air into electrical power, is a major step forward in reducing the effects of carbon emissions from the use of fossil fuels. Although the long-term benefits of harnessing wind power are great, there are some environmental costs that are inevitable including bat fatality.
To minimize bat mortality at wind farms, wind energy developers use a method called curtailment. Curtailment is slowing or shutting down the turbine rotors for a period of time to lessen mortality rates. Stantec, an international services company, is looking to optimize curtailment programs to minimize or avoid turbine-related bat mortality while maximizing wind power efficiency. To create an effective curtailment program, Stantec must understand site-specific information, such as when the bats are active and under what conditions.
Trevor Peterson, a Project Manager at Stantec, is recording bat activity around wind turbines (specifically the top portion, called the nacelle). Peterson is researching how environmental parameters such as temperature and wind speed affect bat activity temporally and seasonally. According to him, “Acoustic monitoring is an ideal method to measure when bats are actually active in the rotor zone of turbines because it is temporally precise. Once you know when bats are active, you can determine the weather conditions associated with higher and lower rates of activity and thereby optimize curtailment to target periods with higher risk. This process, known as smart curtailment or informed curtailment, has allowed a few wind projects to dramatically improve the efficiency of curtailment programs while continuing to avoid most risk to bats.” If it is possible to effectively avoid risk to bats while allowing for substantially more turbine operation, curtailment would be less expensive. Peterson goes on: “I hope that if the cost of curtailment can be driven low enough, the wind industry will be more willing to implement the method broadly.”
Stantec has deployed Wildlife Acoustics Song Meters on ships, offshore wind turbines, land-based wind turbines, meteorological towers, and a variety of other terrestrial habitats from Maine to California. For his work, Trevor Peterson has used Wildlife Acoustics recorders (SM4BAT FS). “I’ve been consistently impressed with [Wildlife Acoustics’] efforts to listen to its customers and update equipment design and options to meet our needs. The world of acoustic bat monitoring is an exciting one to operate in because the technology is improving so quickly in recent years…The omnidirectional mic provides good coverage, and we are able to keep equipment operating for long deployments using external batteries and solar panels.”
Much of Peterson’s work involves differentiating bats from ultrasonic noise generated by wind turbines and other sources. Bat calls recorded at the top of wind turbines or meteorological towers can look different from those recorded near the ground or in more typical foraging habitat, which means Peterson tends to be conservative in assigning calls to groupings sharing similar characteristics (some of which could include multiple species). Peterson explains, “Depending on the dataset, we use a combination of visual analysis and filters (either custom or built into analysis software) to differentiate bats from noise, and then usually a second round of visual analysis to assign passes to a species group based on comparison to reference libraries of known call sequences.”
Peterson uses Kaleidoscope Pro and other tools to view and analyze his calls. However, “[Kaleidoscope] is often my first go-to program to process data.” Kaleidoscope Pro’s bat auto-ID feature is useful to Peterson when it is part of the project’s work scope: “I like to have as much information available to me as possible when analyzing bat data, and the results of auto-ID are useful as a point of reference when reviewing calls.” In more recent years, he has found it to be relevant for much of his work. One particularly useful feature in Kaleidoscope Pro is the ability to create excel spreadsheets that list every recorded call file: “this is very handy when compiling and comparing IDs among methods”
Peterson has developed a workflow that works best for him when analyzing calls. Though it varies from project to project, Peterson generally likes to review his system status files, then review some files in full spectrum, and then again in zero-crossing. This is to ensure that the detectors are working properly. To make troubleshooting easier, a colleague of his created an RMarkdown script to convert SM4BAT status files into a large report summarizing detector function. When applicable, his next step is to process all of his data in Kaleidoscope Pro to convert full-spectrum files into zero-crossing and to perform auto-ID. “If I am doing visual identification, I’ll then compile those results into a separate Excel file. After that, I do everything I can in R to compile and analyze results.” If Peterson is deciding on the presence or absence of a species, he uses a combination of auto-ID and visual methods, depending on the purpose of a project. “If we’re following the USFWS protocols, the process is pretty much following the agency guidance, but visual vetting is a crucial process in determining whether a species is likely present. With acoustics, there’s always a degree of uncertainty in terms of species ID, and I tend to be more conservative in determining presence of species like Indiana bats based on acoustics alone.”
Peterson’s advice for deploying Song Meter users is simple: “Always deploy more equipment than you think you need…equipment can fail for multiple reasons and it’s always good to have backups!”
Keeping track of Stantec’s equipment can also be a challenge. Peterson explains, “staying on top of maintenance (checking microphone conditions, tracking troubleshooting issues) can seem like a full-time job. We’re always working to update our equipment tracking database, but when the field season hits, it can be an all hands-on deck situation!” It is always best to evaluate microphone and clock battery performance as part of general maintenance long before a deployment occurs.
For analyzing bat calls, Peterson finds “it’s important to read as much literature as you can on how echolocation works and understand the degree of overlap in echolocation characteristics between species. Use auto-ID to your advantage, but treat the results with a few grains of salt. Echolocation is an incredibly complex and flexible behavior, and results must be considered in the context of survey design and evaluated with a good understanding of how equipment configuration and deployment may affect recorded echolocation pulses.” As always advised by Wildlife Acoustics, auto-ID results should be manually vetted.
Peterson hopes to eventually “refine analysis of conditions with higher risk at turbine nacelle height to include migratory species and season-specific risk profiles.” Peterson finds that long-distance migratory species typically comprise the majority of acoustic activity recorded at nacelle height and fatalities found during carcass searches. However, policies are often driven by species listed as rare at the state or federal level. This could create an opportunity for more targeted curtailment programs based on site-specific data. Peterson hopes that by replicating his work at multiple projects he will be able to demonstrate the effectiveness of smart curtailment and provide robust data to present to wind companies, agencies, and other stakeholders.
Special thanks to Trevor Peterson at Stantec for providing the content of this case study.