Estimating above-ground biomass of individual trees with terrestrial laser scanner and 3D quantitative structure modelling

  • Tasiyiwa P. Muumbe
  • Paradzayi Tagwireyi
  • Pride Mafuratidze
  • Yousif Hussin
  • Louise van Leeuwen
Keywords: 3D reconstruction, allometric equation, forest carbon, point cloud data, REDD

Abstract

This study explored the feasibility of using the terrestrial laser scanner (TLS) and quantitative structure modelling (QSM) to estimate the above-ground biomass (AGB) of individual trees in the tropical rainforest using data from the Ayer Hitam Forest Reserve, Malaysia. We also tested the influence of varying the runs, the cover set’s diameter and nmin values on the ABG derived. To achieve our objectives, we estimated diameter at breast height (DBH) and height of 100 trees from 26 plots using both TLS and QSM, and field measurements then AGB based on both methods. We observed a powerful positive linear correlation between TLS and field-measured DBH (R2 = 0.97) and between TLS and field-measured height (R2 = 0.77), and a moderately strong relationship between TLS and field-based AGB (R2 = 0.56). TLS and QSM overestimated AGB by 30% of field-based estimates. Varying the number of runs had no significant influence on the AGB derived from the TLS and QSM (one-way ANOVA, p > 0.05) while increasing the cover set diameter led to an increase in the derived AGB (one-way ANOVA, p < 0.05). The QSM was very sensitive to variations in the nmin (one-way ANOVA, p < 0.05). Our study has demonstrated that the TLS and QSM can be used for estimation of AGB for individual trees but with varying reliability. Regardless, our study provides yet another non-destructive approach to the suite of methods for estimating carbon of individual trees for various applications, including Reducing Emissions from Deforestation and Forest Degradation.

Keywords: 3D reconstruction, allometric equation, forest carbon, point cloud data, REDD+

Published
2021-06-25
Section
Articles

Journal Identifiers


eISSN: 2070-2639
print ISSN: 2070-2620