Izvorni znanstveni članak
https://doi.org/https://doi.org/10.5552/crojfe.2026.4340
Evaluating and Modeling of Chainsaw Noise Propagation by Using Artificial Neural Network in Selective Cutting
Sercan Gülci
; Kahramanmaras Sutcu Imam University Faculty of Forestry Department of Forest Engineering Avsar Mah. Bati Cevreyolu Blv. 46040 Onikisubat-Kahramanmaras TÜRKIYE
*
John Sessions
; Oregon State University College of Forestry 336 Peavy Forest Science Complex Corvallis OR 97331-5704 USA
Neşe Gülci
; Kahramanmaras Sutcu Imam University Faculty of Forestry Department of Forest Engineering Avsar Mah. Bati Cevreyolu Blv. 46040 Onikisubat-Kahramanmaras TÜRKIYE
* Dopisni autor.
Sažetak
The investigation of the effects of the noise generated by harvesting equipment on the environment is one of the important topics in sustainable forestry. During timber harvesting, not only workers but also wildlife are exposed to the noise generated. Exposure to noise has both direct and indirect effects on humans and wildlife. The negative effects of noise exposure can be observed depending on its intensity and duration. Noise exposure, which has various psychological and physiological effects on humans, also negatively affects plants and animals. In this study, sound measurements of the chainsaw were conducted during thinning operations within the boundaries of the Alara Forest Management Directorate in Alanya, Antalya Province. The measurement area is a Turkish red pine (Pinus brutia Ten.) stand with a canopy density of 60–65%, a slope of 30–35%, and tree diameters ranging from 20 to 35 centimeters. The noise emitted by the chainsaw during production, ranging from approximately 1 meter to 200 meters, has been modeled using a feedforward Artificial Neural Network (FANN) for sound propagation. The measurement data was used 60% for training, 20% for testing, and 20% for validation. Random trees were assigned to noise attenuation effects on the sound according to the stand characteristics of the study area. Thus, it was aimed to create a realistic sound propagation model and estimation maps. The average performance metrics of the model, RMSE and R² values, were calculated as 4.84 and 0.88, respectively. According to the sound propagation model predicted by the FANN as a static model, it is estimated that the distance at which the chainsaw could affect wildlife behavior is 400 meters or less.
Ključne riječi
forest operations, sound effect, machine learning, wildlife, habitat
Hrčak ID:
343133
URI
Datum izdavanja:
16.1.2026.
Posjeta: 138 *