High-Resolution Optical-Zoom Drone Imagery for the First Quantitative Estimate of Euphorbia dendroides in a Coastal Cliff Sector of Monte Conero

Coastal cliffs remain among the least investigated terrestrial environments because steep slopes, instability and limited accessibility strongly constrain conventional vegetation surveys. As a consequence, ecologically and biogeographically relevant plant populations may be known only through presence records, with little quantitative information on their abundance and spatial distribution. This is the case of Euphorbia dendroides L. at Monte Conero, where the species occurs near the northern limit of its Adriatic distribution range on largely inaccessible calcareous cliffs.

This study provides the first quantitative estimate of this population by combining high-resolution optical-zoom drone imagery with a two-phase sampling design over a 23.78 ha coastal cliff sector. In the first phase, the area was divided into 25 × 25 m cells and screened to identify those potentially suitable for E. dendroides. In the second phase, sampling plots were selected within these cells to obtain a spatially distributed sample used for population estimation. Nadir images were acquired from 90 m using a 16× optical zoom (405 mm equivalent focal length), with a ground sampling distance of approximately 1 mm pixel. Each analysed image represented 44.3 m², allowing individual plants to be visually identified and counted.

A total of 206 images covering 9,126 m² were analysed. E. dendroides was detected in 80 plots, with 420 individuals directly counted. Using the two-phase sampling estimator, the population within the entire study area was estimated at 9,154 individuals (RSE = 14.9%; sampling error = ±1,364 individuals). Higher abundances occurred in the northern sector and closer to the coastline.

These results show that high-resolution optical-zoom drone imagery combined with targeted two-phase sampling enables individual-based population estimates in steep and inaccessible environments. The approach is repeatable, spatially explicit and potentially transferable to other plant species recognizable from high-resolution imagery, providing a flexible framework for biodiversity monitoring and future change detection.