Epsilon Archive for Student Projects

Abstract

Accurate biomass and leaf area index (LAI) estimation are crucial for understanding forest productivity and carbon dynamics. Models are crucial to estimating forest biomass and LAI. This study aimed to estimate site-specific aboveground biomass and LAI in a young Scots pine stand in southern Sweden. Destructive sampling was carried out to obtain data for biomass and LAI. Site-specific models were developed, including an individual tree LA model based on diameter at breast height (DBH), LAI models based on stand density (LAImodel1) and basal area (LAImodel2), stand-level biomass models developed based on stand density and basal area as independent variables, and tree-level biomass models developed based on DBH. Directly measured LAI values were used to validate a generalized LICOR model for indirect LAI estimation, which was compared to the site-specific LAI model developed in this study. Evaluation metrics such as mean absolute error (MAE), mean squared error (MSE), and root mean squared error (RMSE) were used to compare the site-specific and generalized model. Results showed that site-specific LAI models of this study gave reasonable estimates (LAImodel1 have MAE = 0.166, MSE = 0.039, RMSE = 0.198, and LAImodel2 – MAE = 0.011, MSE = 0.000, RMSE = 0.015) compared to the generalized model (MAE = 0.66, MSE = 0.435, RMSE = 0.66). Directly measured LAI and indirectly measured LAI using the generalized LICOR model had a weak correlation (R2 = 0.1339). For stand-level biomass models, stand density showed moderate correlation (R2 = 0.49) to the total stand biomass, and basal area was strong correlation (R2 = 0.75) with the total stand biomass. These findings demonstrate the importance of using site-specific models to reduce prediction errors when estimating biomass and LAI in forest stands.