Sound level measurements from audio recordings provide objective distance estimates for distance sampling wildlife populations
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Distance sampling is widely used to estimate animal population densities by accounting for imperfect detection of individuals with increasing distance from an observer. Distance sampling assumes that distances are measured without error; however, it is often applied to human estimated distances, which are known to be inconsistent, inaccurate, and biased. We present an objective technique for estimating distance to vocalizing individuals that relies on the relative sound level ( RSL ) of the vocalization extracted from autonomous recording unit ( ARU ) recordings and show the error is less than human estimated error extracted from a literature case study. RSL predicted distances can be obtained by manual measurement in sound viewing software, or automatically with automated signal recognition software. We built calibration datasets of Ovenbirds ( Seiurus aurocapilla ) and Common Nighthawks ( Chordeiles minor ) recorded at known distances and used regression of RSL from those recordings to predict distance. There was no error bias of RSL predicted distances when compared to known distances for Common Nighthawk, minimal error bias for Ovenbird, and error from all RSL predicted distances was less than human estimated error extracted from the literature. We then simulated ARU point count surveys with a known density and estimated that density with distance sampling to test whether RSL distance prediction does not violate the assumption that distances are measured without error. There was no difference in density estimates from known distance and density estimates obtained from RSL predicted distance, while density estimates contaminated with human estimated error were significantly lower than density estimates from known distance. We found that a calibration dataset of approximately 300 vocalizations was suitable to minimize error for both species, and so conclude that RSL distance prediction is an accessible method of improving distance estimates relative to human estimation. We provide general recommendations on how to collect calibration recordings for the application of RSL distance prediction to other species and areas.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it