Hi everyone! I’m new in Motus so I have some ideas I want to try with my data. I have 10 individuals ranging from 3 to 130 days of data, and high residence in a station. I’ve chechked some tutorials, and some people uses short period time, in days, to pressume behavior, like resting (with no difference in signal, like a flat line in a graphic with time and signal strenght), or fly-by (with a classic pattern of a inverse Gaussian bell curve).
Has someone used signal strenght pattern, between records, as a proxy of behavior? Do you think it’s possible? I’m working with shorebirds, so I was thinking any idea from passerines would work.
My advisor Dr. Mike Ward and I have been working on a method to convert Motus data to activity patterns. We currently have a paper in review, but in the mean time I have also been working on a GitHub where I provide code to (1) download Motus data and seperate them into individual tag files, (2) calculate thresholds during periods of inactivity (for our study species, Wood Thrushes, we used night data, but this can easily be changed), and then (3) use those thresholds to estimate activity ( GitHub - wardlab-uiuc/Motus_Telemetry_Activity_Pipeline: R pipeline for estimating movement-based activity from Motus automated radiotelemetry detections using receiver-type thresholds, proportional signal change, and diagnostic screening. · GitHub ). I also tried to make the ReadMe comprehensively explain how this method calculates thresholds and handles receiver noise and the log signal scale. It results in an hourly activity output (proportion of active detections divided by the proportion of total detections), but it does not infer specific behavior. I would be happy to meet with you to further discuss this method if it is of interest to you! My email is lbrunk2@illinois.edu.
We have used a CTT node placed near a nest to infer behavior using signal strength patterns to decide whether birds are incubating, feeding nestlings, and tending fledgelings. Not published, but it seems from visual verification to be a reliable indicator of behavior.
We just submitted revisions on a paper addressing this and have made the paper available on BioRxiv: roostR: An R package to examine diel activity patterns from Motus radio telemetry data
You can install roostR from github. While the package was developed (so far) for overnight activity, the package estimates activity throughout a 24 hour period. You can use the package to get the rolling median then plot. You could decide on a threshold and label as active or inactive.