Using Drone-Based Photogrammetry to Detect Ecosystem Change in the Schooner Cove Sand Dunes, Pacific Rim National Park Reserve

In collaboration with Pacific Rim National Park Reserve Parks Canada Under Supervision of Geomatics Coordinator Mike Collyer

Abstract

Since 2012 Parks Canada has been working with the Tla-o-qui-aht First Nation, academic partners and volunteers to restore the ecological integrity and recover associated species at risk at the Tayus (Schooner Cove) dune in the Pacific Rim National Park Reserve and the traditional territory of the Tla-o-qui-aht First Nation.  As with most coastal sand ecosystems on the Canadian and American Pacific coasts, the Tayus dunes were degraded by the establishment of invasive grasses (Ammophila arenaria and breviligulata).  These grass species both out compete native plant species and interfere with natural sand transport regime, stabilizing the dune system and allowing accelerated succession to a forest ecotope.² By helping to conserve biodiversity through the restoration and preservation of disturbance-driven habitats, the project supports COSEWIC- and SARA-listed species of conservation concern, such as the endangered Pink Sand-verbena (Abronia umbellata) and the threatened Silky Beach Pea (Lathyrus littoralis), that depend on these conditions.¹⁰

Parks Canada has been monitoring and assessing the progress of this work by implementing high accuracy land surveys using total stations, RTK GPS and drones to track changes in vegetation and geomorphology. However, Parks Canada has not to date, fully analyzed the drone imagery. This project aims to further inform the restoration of healthy and dynamic sand-transportation processes by evaluating how effectively photogrammetric analysis of drone imagery can detect changes in coastal dune environments. Using a consistent set of ground control points (GCPs), resulting in orthomosaics, digital surface model (DSM) difference maps, and elevation transects, this work will assess sand distribution, dune dynamics, and associated vegetation change across multiple years (December 2020 to January 2025), while also considering factors that influence DSM precision and accuracy.

Between December 2022 and January 2025, Kinnikinnick expanded 35.75 m² (net amount) along the landward dune margin, particularly in the middle section, correlating with sand loss, longer transport distances, and low elevation variability close to the margin in adjacent transects, indicating reduced backdune sand movement and limited sediment inflow from the foredune and backshore. Difference models and elevation profile comparisons show that the Schooner Cove dunes are undergoing active dune-building, with a dynamic nature and widespread sand deposition - especially in the foredune, where accretion reached 0.61 to 0.77 m - and only localized areas of minor elevation loss in the backdune. While model comparability generally fell within acceptable accuracy ranges for the 2021-2025 datasets, discrepancies were pronounced when comparing models generated with different drone systems, especially the 2020 Mavic 2 Pro dataset, which showed vertical GCP RMSEs of 4.4-4.7 cm and horizontal errors of 5.3-7.3 cm as well as with 0.825 px the overall model reprojection error. These uncertainties must therefore be considered when interpreting small elevation changes or mapping vegetation encroachment. Overall, the photogrammetric outputs demonstrate strong comparability between models and provide a reliable basis for detecting broader, longer-term geomorphological trends, though they are less suited for identifying very small or short-term changes, some of which may reflect brief weather events. To improve future assessments, surveys should be conducted more frequently, with consistent timing, standardized survey parameters, and the use of the same or comparable drone systems and fixed checkpoints.