Identification of 20 species from the Peruvian Amazon tropical forest by the wood macroscopic features

ABSTRACT

Background:

The biodiversity of the Peruvian Amazon tropical forests is one of the most expressive in the world, with 2500 forest species, although restricted to about 250 tree species used commercially. This species diversity indicates the challenge magnitude for research in taxonomy and timber species identification. Likewise, it implies the complexity of biodiversity conservation and restoration measures, which are directly related to the control of deforestation, cutting and transport of illegal wood. With this objective, the present study describes the macroscopic wood anatomical features in order to identify 20 tree species from Peruvian Amazon Forest, “Selva Central”, including an identification key and tree species botanical validation.

Results:

The wood species are included in 12 families commonly found in the tropical forests of Peru, highlighting the Fabaceae (25%), Moraceae (15%), Podocarpaceae and Lauraceae (10%) families and are sold as timber for several uses and applications in the internal market and for export. The wood species presented common anatomical features, such as diffuse porosity, visible axial parenchyma mainly distinct, and, eventually, with ray storied, e. g: A. cearensis, M. balsamum and M. peruiferum.

Conclusion:

The tropical tree species identification is possible by analyzing their wood macroscopic anatomical structure. The results can be also applied in the wood trade traceability by controlling deforestation and illegal wood commerce and in proposing policies for biodiversity conservation and sustainable use of natural resources. They constitute, likewise, a database for the recent wood identification methodologies presented in the specialized literature.

Key words:
Forensic wood identification; wood anatomy identification; Tropical timber species; Peruvian Amazon forests; misidentification; illegal logging

HIGHLIGHTS

Identification by anatomical structure enable wood trade traceability. Tree species identification is essential for controlling deforestation. Machine learning satisfactorily works for wood anatomy. Convolutional neural networks can identify tropical Peruvian Amazon species.