Drone-calibrated acoustic footprints infer group size in franciscana dolphins Pontoporia blainvillei

Conventional visual surveys of the cryptic franciscana dolphin Pontoporia blainvillei are constrained by turbid near-shore waters and the species’ tendency to cluster in small groups. We combined passive acoustic monitoring (PAM) with drone counts to calibrate an acoustic footprint detection model and infer group size from acoustic data. Click trains were detected in PAMGuard, localized, and clustered with HDBSCAN into click-train clusters (CTCs). Drones provided visual group sizes during synchronized sections. A 2-step spatiotemporal matching paired CTCs to visual groups. An independent manual audit estimated a 6.4% false-positive rate, used to correct automated counts. For paired clusters, we fit GLMs with an offset for animal-seconds and alternative distance terms; AIC selected a linear negative binomial model. The calibrated baseline acoustic footprint at the trackline was r 0 = 0.028 click trains animal -1 s -1 (95% CI: 0.014-0.056; CV = 0.34; R 2 = 0.54), equivalent to 1.68 (CI: 0.84-3.36) min - 1 . We obtained 28 paired CTCs from 8 visual groups and applied the calibration to 267 unpaired CTCs. Predicted group sizes had a median of 1.34 individuals (95% CI: 0.79-2.57; CV = 0.15), consistent with predominantly small units. This framework provides a pathway from clicks to counts and delivers an important step required for passive acoustic density estimation of this species.