Limitations of forest structural complexity indices as indicators of old-growth conditions in coastal temperate rainforests

Airborne Laser Scanning (ALS)-derived forest structural complexity (FSC) relies on the assumption that structurally complex forests exhibit greater three-dimensional heterogeneity in LiDAR point clouds than simpler stands. Although FSC is often assumed to increase with forest succession and serve as a proxy for old-growth conditions, it is also shaped by site quality, topography, climate, and disturbance. Here, we tested whether old-growth forests exhibit greater FSC than earlier successional stages. We assessed six ALS-derived FSC indices (e.g., canopy rugosity, canopy entropy, foliage diversity) against 13 old-growth structural attributes (e.g., stand age, biomass, vertical complexity, top height) measured in 324 plots across Vancouver Island, British Columbia, Canada (inventoried 1990 – 2019). Using random forest models with five-fold cross-validation, we evaluated FSC index performance and redundancy, and then examined FSC responses across gradients of stand age, forest maturity (IMAT; an aggregate measure of old-growth structural attributes excluding age), and site productivity (site index; SI, defined as expected tree height at 50 years). Canopy entropy, canopy rugosity, and median absolute deviation of height were highly redundant. However, canopy entropy showed the strongest relationship with old-growth structural attributes (R² = 0.57 ± 0.08) and was selected as the FSC index for subsequent analyses. Using canopy entropy as the FSC index, relationships with top height and live tree volume were asymptotic, with FSC saturating at approximately 45 m and 900 m³ ha⁻¹ , respectively. In contrast, FSC exhibited a quadratic relationship with DBH and age, increasing up to ∼250 cm DBH and ∼210 years, followed by a slight decline. FSC declined in older forests on low-productivity sites (SI < 20 m) but remained high in productive old-growth forests (SI > 20 m). FSC was more strongly related to IMAT than to age alone, yet also showed clear saturation. Together, these results indicate that FSC primarily reflects biomass accumulation and site productivity rather than successional status, making it a robust descriptor of forest development but an insufficient proxy for old-growth forests.