Advanced Journal of Environmental Science and Technology

ISSN 2756-3251

Advanced Journal of Environmental Science and Technology | Vol. 17 (6), pp. 1247-1260, June, 2026

Research Article

Machine-Learning-Guided Acoustic Monitoring Quantifies Coral Reef Biodiversity Recovery Following Artificial Reef Deployment

kieran Murphy¹*, Siti Aisyah², Mateus Costa³

¹School of Biological, Earth and Environmental Sciences, University College Cork, Cork, Ireland
²Institute of Ocean and Earth Sciences, Universiti Malaya, Kuala Lumpur, Malaysia
³Department of Oceanography, Federal University of Pernambuco, Recife, Brazil

Corresponding Author: Kieran Murphy — [email protected]

Abstract

Visual SCUBA surveys of coral reef restoration sites are labor-intensive, restricted to daylight hours, and miss cryptic nocturnal reef fauna. We deployed autonomous hydrophone arrays paired with a convolutional neural network trained on 4,200 hours of marine soundscapes across degraded and restored tropical reefs. Acoustic complexity and snapping-shrimp pulse rates correlated strongly (R² = 0.91) with fish species richness and live coral cover within 18 months of modular mineral-accretion reef installation. The non-invasive framework detected larval fish settlement pulses three months before they became detectable via conventional visual transects.

Keywords: passive acoustic monitoring; coral reef restoration; marine soundscapes; machine learning; biodiversity assessment; artificial reefs

Manuscript Timeline: Received 19 May, 2026; Revised 16 June, 2026; Accepted 28 June, 2026; Published 30 June, 2026.

Citation: Murphy, K., Aisyah, S., and Costa, M. (2026). Machine-Learning-Guided Acoustic Monitoring Quantifies Coral Reef Biodiversity Recovery Following Artificial Reef Deployment. Advanced Journal of Environmental Science and Technology, 17(6), pp. 1247-1260.