NEXUS – The Acoustic Bubble: Interim Report (v1.1)

Stationary ultrasonic detectors are routinely used in bat research, yet the influence of microclimatic conditions on detection range has not been systematically quantified. This interim report presents empirical evidence for the "acoustic bubble"—a weather-dependent detection zone around stationary ultrasonic sensors caused by atmospheric attenuation. Over 20 valid recording nights at 10 sites in the Paderborn region (February–July 2026; 22 of 30 planned nights completed), a self-contained microclimate monitoring system (NEXUS: Seeed XIAO ESP32S3 Sense, BME680, AIR530, SparkFun SEN-15901) was operated alongside a TeensyBat T4.1 ultrasonic detector. A within-session fixed-effects approach—controlling for habitat-specific confounders—demonstrates a significant relationship between the theoretical bubble radius and the detection rate (Pearson r=0.592, R²=0.351, p<0.0001, n=79 bins, 20 nights); this result is robust according to a leave-one-session-out jackknife analysis (coefficient range: 1.76–2.14). Species-specific bubble radii align with attenuation laws defined by ISO 9613-1. Bioacoustic species identification was performed using dual-CNN cross-validation (BatDetect2 + BattyBirdNET, confidence ≥ 0.85), confirming five species as the "gold standard" (560 detections, 10 sites). A full analysis incorporating the remaining ten recording nights will be presented in the final version. v1.1 – Data correction and completion This version corrects and completes the dataset underlying v1.0: Full inclusion of the June 26, 2026 session in the fixed-effects analysis (erroneously excluded from the processing pipeline in v1.0); dataset expanded from 19 to 20 valid nights (n=79 instead of n=73 bins) Updated core result: Pearson r=0.592, R²=0.351, p<0.0001 (previously r=0.621, R²=0.386) Consistent correction of night counts and date range throughout the manuscript (22 total nights, 20 valid, February–July 2026) Correction of the raw (non-session-normalized) cross-habitat correlation to the actually reproducible value (r=−0.080, p=0.481) Language precision improvements for overstated claims ("proves" → "provides evidence for") Addition of a previously unreported Moon Illuminance Index finding (significant correlation with activity, but in the direction opposite to the Lunar Phobia hypothesis) for full transparency Clarification in Section 4.5 that the ISO 9613-1 example calculation uses a theoretical reference condition, not the dataset-wide mean The core scientific finding (Acoustic Bubble hypothesis, within-session fixed-effects approach) remains unchanged and robustly confirmed.