Estimating tree bole volume using artificial neural network models for four species in Turkey

Area of Science:

• Forestry
• Computational modeling
• Machine learning

Background:

• Accurate tree bole volume estimation is crucial for forest management and resource assessment.
• Traditional methods for tree volume estimation have limitations in capturing complex relationships.
• Artificial Neural Networks (ANNs) offer potential for improved accuracy in modeling nonlinear relationships.

Purpose of the Study:

• To evaluate the effectiveness of Artificial Neural Network (ANN) models for estimating tree bole volume.
• To compare ANN models (Back Propagation and Cascade Correlation) against traditional methods.
• To determine the most reliable ANN architecture for tree volume estimation across multiple species.

Main Methods:

• Artificial Neural Network (ANN) models, specifically Back Propagation (BPANN) and Cascade Correlation (CCANN), were developed.
• Tree bole volumes were estimated for Scots pine, Brutian pine, Cilicica fir, and Cedar of Lebanon.
• ANN model performance was compared against the centroid method, taper equations, and standard volume tables.
• Validation was performed using Smalian's formula on measured tree bole sections.

Main Results:

• Cascade Correlation Artificial Neural Network (CCANN) models demonstrated high reliability for tree bole volume estimation.
• CCANN models provided unbiased results and exhibited lower percentage errors compared to most traditional methods.
• The study highlights the superiority of CCANN in capturing complex nonlinearities inherent in tree volume modeling.

Conclusions:

• Cascade Correlation Artificial Neural Network (CCANN) models are recommended for accurate tree bole volume estimation.
• ANNs, particularly CCANN, offer significant advantages over traditional methods for tree volume assessment.
• The findings support the use of advanced computational techniques in forestry for improved resource management.