Soil and Water Research - In Press

Linking seasonal fractional vegetation cover dynamics with soil organic carbon stock and microbial indicators in tropical agroecosystemsOriginal Paper

Ni Made Trigunasih, Moh Saifulloh, Ida Bagus Putu Bhayunagiri, Zulkarnain Zulkarnain

Reliable indicators of early soil biological change remain limited in tropical agroecosystems, where Soil Organic Carbon (SOC) stocks may respond more slowly than microbial processes. We evaluated whether seasonal vegetation dynamics derived from Sentinel-2 fractional vegetation cover (FVC) are associated with spatial variation in SOC stock and microbial indicators in Jembrana, Bali, Indonesia. We mapped seasonal FVC from 2019 to 2024 and derived site-level metrics of mean cover and temporal variability (standard deviation, anomaly, coefficient of variation, and a temporal stability index). In July 2023, we sampled topsoil (0–30 cm) at 12 sites representing contrasting land uses and topographic settings. We calculated SOC stock from organic carbon concentration, bulk density, and sampling depth, and measured basal respiration and culturable microbial density (colony-forming units, CFU). Vegetation cover peaked consistently during the wet season (December to February), and mean site FVC ranged from 0.31 to 0.99. Mean FVC showed positive but non-significant associations with culturable microbial density (Spearman’s ρ = 0.48, P = 0.114) and basal respiration (ρ = 0.29, P = 0.361), whereas higher vegetation variability metrics tended to coincide with lower culturable microbial density (ρ = -0.43 to –0.51, P = 0.090 to 0.163). SOC stock showed near-zero coefficients and no statistical evidence of association with vegetation metrics (ρ = 0.09, P = 0.781) or microbial indicators (ρ = 0.01, P = 0.975). Principal component analysis of FVC traits explained 99.65% of the variance and separated sites along a gradient from stable, high cover to more variable, lower cover. Overall, FVC stability metrics captured spatial differences that were directionally consistent with microbial indicators, but associations were not statistically significant in this dataset (n = 12). Larger, replicated studies with repeated soil sampling are required to evaluate whether seasonal FVC metrics have robust predictive utility for SOC stock and soil biological indicators.