Evaluating the growth performance of a newly planted forest using multi-temporal UAV imagery
摘要
Estimating tree height growth and survival rate are important for characterizing forest growth performance and evaluating dynamic changes during the early development of newly planted forests. This study evaluated tree height growth and survival rates in a newly forested area of Chinese fir (Cunninghamia lanceolata (Lamb) Hook) and Schima superba Gardner & Champ. planting from two temporal Unmanned Aerial Vehicle (UAV) images over a four-year period. Individual tree crowns were delineated by visual interpretation, and tree heights were estimated using the local maxima method. Using the 2023 digital surface model (DSM) combined with the 2019 digital terrain model (DTM), UAV-based tree height estimation achieved high accuracy, with a root mean square error (RMSE) of 0.48 m for Chinese fir and 0.37 m for S. superba compared with field survey measurements in 2023. It reduced RMSE by 0.30 m for Chinese fir and 0.50 m for S. superba compared to using the 2023 DTM for ground height estimates. From 2019 to 2023, the survival rates of Chinese fir and S. superba were 92.39% (1,687/1,826) and 73.41% (392/534), respectively. These dead trees had average heights of 1.51 m for Chinese fir and 1.56 m for S. superba in 2019. The average tree heights of Chinese fir and S. superba in 2023 were 7.83 m and 5.78 m, respectively. The average tree height growth of Chinese fir and S. superba was 5.50 m and 4.03 m between 2019 and 2023, respectively. This study highlights the feasibility and effectiveness of using multi-temporal UAV imagery for long-term, accurate monitoring of newly planted forests.