Investigation of changes in leaf area ındex in different forest stands.

Leaf area ındex (LAI) is a fundamental metric of forest canopy structure, driving photosynthetic capacity, carbon sequestration, and ecosystem productivity. In this study, we quantified LAI variation across six forest stand types (pure fir, pure beech, pure Scots pine, mixed coniferous, mixed deciduous, and mixed deciduous-coniferous) and six developmental stages in the Düzce region of Turkey, an area characterized by mixed, heterogeneous woodlands. Field measurements from 260 systematically distributed sample plots collected in 2015 employed hemispherical photography for LAI, the N-tree method for basal area (BA) and diameter at breast height (DBH), and atmospherically corrected Landsat 8 OLI imagery for NDVI. Descriptive analyses revealed LAI values ranging from 0.36 m/m in pure Scots pine to 6.30 m/m in mixed deciduous stands. One-way ANOVA and Duncan's multiple-range tests confirmed significant differences among stand types (p < 0.01), with mixed and vertically stratified stands exhibiting the highest LAI. Developmental stages showed increasing mean LAI trends from juvenile (ab) to mature (d) classes, though stage-only ANOVA was not significant (p = 0.378) due to high within-stage variability (CV ≈ 31%). Pearson correlations indicated moderate positive relationships between LAI and both DBH (r = 0.49) and BA (r = 0.53), whereas NDVI displayed the strongest association (r = 0.75 overall; up to r = 0.80 in mixed stands). A multiple linear regression model integrating NDVI, DBH, and BA explained 60.6% of LAI variance (F = 129.7, p < 0.001; adjusted R = 0.601), with NDVI emerging as the dominant predictor (standardized β = 0.683), followed by DBH (β = 0.326) and BA (β = 0.187). These findings underscore the complementary value of integrating spectral indices and structural parameters in the estimation of LAI, particularly in heterogeneous forest stands. The structural complexity of mixed stands appears to play a critical role in enhancing canopy development. To improve estimation accuracy in conifer-dominated or high-LAI forests, future studies should consider incorporating alternative vegetation indices and LiDAR-derived structural metrics to overcome limitations such as spectral saturation and insufficient vertical resolution. Such integrated approaches can significantly enhance the scalability and cost-effectiveness of forest health and productivity monitoring efforts.