Forest Leaf Area Index (LAI) Estimation Using Airborne Discrete‐Return Lidar Data

Abstract

AbstractThe Leaf Area Index (LAI) is one of the most important parameters, which controls biological and physical processes associated with vegetation on the Earth's surface, such as photosynthesis, respiration, transpiration, carbon and nutrient cycles, and rainfall interception. Therefore, rapid, reliable and objective estimations of LAI are essential. In this study, we used a simplified approach to compute laser penetration index (LPI) from airborne discrete‐return LiDAR data, and LPI was computed based on corrected echo intensity for the first time. Using the variable of LPI, we built an LAI estimation model based on Beer‐Lambert law. This approach was applied to a forest area in Dayakou, Gansu Province of China. The accuracy of the corrected intensity‐derived LAI estimation model was compared with that of uncorrected intensity‐derived and echo counts‐derived model. The results show that the corrected echoes intensity can improve the accuracy of LAI estimation. To assess validity and generalization of the model, we validated the optimum model using the Leave‐One‐Out Cross‐Validation (LOOCV) procedure, and show that the model has no overfitting and is more general. Finally, we examined the accuracy of predicted LAIs with 16 field‐measured LAIs which were not involved in the modeling process and indicated that LAI can be estimated with a high accuracy in mountains area by corrected echo intensity. The comparison indicates the LiDAR‐derived LAI (R2=0.825, RMSE=0.165) is much higher than that of the LAI from Landsat TM images (R2=0.605, RMSE=0.257). It means that airborne LiDAR data can be used to obtain high‐accuracy LAI estimation and provide reliable data for ecological environment research.