Abstract: Leaf area index (LAI) is a critical parameter for characterizing canopy structures. The spatiotemporal patterns of LAI reflect the multi-scale adaptive mechanisms of forests, spanning from individual leaf expansion and population structure optimization to coordinated ecosystem functionality. However, research on the spatiotemporal dynamics of LAI remains limited, particularly in karst primary forests that feature complex community structures, high rock exposure rates, and heterogeneous soil distribution patterns. Our research focused on a 1.28 ha karst forest plot in southwest China, where we quantified LAI seasonality using the LAI-2200 analyzer during four critical phenophases: January (dormant period), April (growth period), July (maturation period) and October (senescence period). Structural equation modeling (SEM), variance partitioning and hierarchical partitioning analysis were employed to quantify the relative contributions of biological, abiotic and spatial factors (represent community spatial connectivity shaped by ecological processes including seed dispersal, resource distribution and competition) to LAI variation. The results revealed distinct seasonal fluctuations in LAI, with significant spatial autocorrelation occurring at characteristic scales of 11.4 m (dormant period), 20.1 m (growth period), 66.3 m (maturation period), and 19.2 m (senescence period), respectively. Spatial factors were the dominant factors in explaining LAI variation across growth, maturation and maturation periods, accounting for 32% to 80% of the explained variance. During the dormancy period, however, the proportion of deciduous trees was the most important factor (32%). The spatial factors exhibited a positive correlation with LAI during the growth and dormancy periods but showed an inverse relationship in other periods. Biological drivers differentially regulated LAI across growth, senescence and dormancy period. Tree height and deciduous species proportion reduced LAI, whereas crown width, density of small trees and stand density increased it. Potassium (inhibiting) and phosphorus (promoting) elements were the primary soil elements influencing LAI. Topographic factors (altitude, slope position and slope aspect) affected LAI through direct, indirect, or combined effects, with slope aspect being the dominant topographic factor. In summary, spatial distribution of LAI exhibited seasonal dependence, which was influenced by spatial, stand, soil and topographic factors, etc. This study identified the primary drivers of LAI in karst evergreen and deciduous broadleaf forests during different seasons, and simultaneously demonstrated the importance of long-term forest monitoring and targeted management strategies, while highlighting the necessity of incorporating spatial factors into LAI estimation and modeling frameworks.

Key words: Karst forest, Leaf area index, Spatial autocorrelation, Structural equation modeling