Height-threshold-dependent estimates of forest gap characteristics, spatial patterns, and dynamics: Evidence from multi-temporal UAV–LiDAR

Forest gaps record disturbance, regeneration, and canopy structural change, yet estimates of gap closure depend strongly on the height threshold used to define when a gap is considered closed. This issue is particularly relevant for three-dimensional LiDAR analyses, where gap structure and dynamics can be quantified across vertical canopy layers. We used multi-temporal UAV–LiDAR data from a temperate forest in northeastern China to test how six fixed height thresholds (2–12 m) influence estimates of gap characteristics, spatial patterns, and dynamics. Gap density and gap fraction increased consistently with increasing height threshold, indicating that lower thresholds identified smaller subsets of forest gaps, whereas higher thresholds captured broader ones. Gap-size frequency distributions were consistent with power-law scaling across thresholds and years, with exponents of 3.38–3.74 in 2018 and 3.27–4.01 in 2023. Gap distributions were generally clustered, but spatial aggregation weakened with increasing height threshold. Gap dynamics were also threshold-dependent. Gap formation and closure rates increased from the lower thresholds to 10 m, whereas area closure rate generally increased with increasing height threshold. Strong covariation among height threshold, gap density, gap fraction, and gap disappearance rate (r > 0.90) indicated that threshold effects were systematic and coordinated across gap metrics. Therefore, height thresholds shape not only numerical estimates of gap metrics, but also whether a canopy opening is interpreted as open, partially filled, or closed. Empirical relationships among thresholds may support cautious cross-study comparisons, but only when differences in gap-closure criteria are explicitly considered.