Evaluating terrestrial laser scanning for structural characterization of mangrove forests in Southeastern Brazil

Abstract

Mangrove forests present significant structural variations at different spatial and temporal scales, reflecting complex interactions between climatic, hydrological and geomorphological factors. The analysis of these parameters by traditional methods has allowed for the description of the organization, architecture and distribution of species in mangrove ecosystems. In recent years, the use of Terrestrial Laser Scanning (TLS) has been consolidated as an innovative and efficient tool for forest structure studies, standing out for its speed and accuracy. This study aimed to compare the performance of TLS with traditional methods concerning data collection to calculate structural parameters in mangrove forests in southeastern Brazil. The Mean Absolute Percentage Error (MAPE) was 7.39 % for mean diameter at breast height (DBH), 13.92 % for mean height, 4.74 % for density and 12.58 % for basal area. The Root Mean Square Error (RMSE) was of 8.92 % for mean DBH, 19.57 % for mean height, 6.75 % for density and 15.2 % for basal area. Regression analyses between the methods indicated high correlations, with coefficients of 0.78 (mean DBH), 0.98 (mean height), 0.95 (density) and 0.82 (basal area). These findings point to the potential of using TLS as an effective alternative for the structural characterization of mangrove forests. This method provides significant gains in terms of time savings, increased quality of the collected information and detailed data recording, allowing for its preservation for different analyses in future studies.