Integrated terrestrial and airborne LiDAR systems to monitor stand structure variations in dryland forests affected by thinning treatments

Light Detection and Ranging (LiDAR) technologies have become useful tools for forest monitoring, enabling precise evaluation of structural attributes that inform management decisions. However, the potential of integrating Mobile LiDAR Scanning (MLS) and Airborne LiDAR Scanning (ALS) for monitoring forest structure across stands with varying management remains underexplored. This study assessed the capabilities of MLS, ALS, and their fusion for quantifying tree- and stand-level structural characteristics in two dryland pine forests in Israel: HaKedoshim (semi-arid) and Yatir (arid). In HaKedoshim, both MLS and ALS data were collected and integrated; in Yatir, ALS alone was used. ALS-MLS fusion models demonstrated strong agreement with traditional field inventory measurements, achieving high correlations for diameter at breast height (DBH, R² = 0.88), stem basal area (BA, R² = 0.86), crown projection area (CP, R² = 0.88), and canopy volume (CV, R² = 0.85). Stand-level attributes such as tree density (TD, R² = 0.98), average tree canopy projection (R² = 0.84), and stand canopy cover (CC, R² = 0.91) were also reliably estimated. However, canopy top height (CTH) was predicted with lower precision (R² = 0.68), reflecting challenges in vertical segmentation using LiDAR and field measurement errors. As detected by the ALS-MLS models, thinning treatments reduced TD and CC at the stand level while average tree CP and CV increased. ALS-derived estimates of Plant Area Index (PAI) demonstrated high accuracy for understory (R² = 0.82), overstory (R² = 0.90), and ecosystem PAI (R²= 0.85). PAI_Ecosystem values ranged from 1.57 to 3.22 m²/m² in HaKedoshim and from 0.65 to 0.98 m²/m² in Yatir, highlighting the role of climatic aridity in shaping forest structure. The analysis of vertical PAI profiles from ALS data revealed that thinning treatments (applied 10 years ago) consistently reduced overstory PAI, while understory PAI increased due to thinning only in the more humid HaKedoshim site. Overall, the MLS–ALS fusion approach enhanced multi-scale assessments of forest structural properties. Our results offer a scalable framework for monitoring forest structure, including vertical canopy partitioning as affected by climate and thinning treatments, with direct implications for dryland forest management and ecosystem modeling.