디지털 토양도 작성 기법을 적용한 산림 토양 탄소 저장량 평가 - 편마암 지역을 대상으로 - (Assessment of Forest Soil Carbon Stocks Using Digital Soil Mapping in a Gneiss Area)

서지정보

· 발행기관 : 한국지형학회
· 수록지 정보 : 한국지형학회지 / 31권 / 3호 / 29 ~ 46페이지
· 저자명 : 정관용, 심우진, 이종명, 지동훈

초록

This study aims to evaluate carbon stocks at varying soil depths in the forest soils of the Namyangju region using digital soil mapping techniques. The forest soils in the study area were classified into organic and mineral layers, and carbon stocks were assessed for each layer to analyze their correlation with environmental variables. The random forest algorithm was employed to predict the spatial distribution of forest soil carbon stocks. The analysis of forest soil carbon stocks in the study area indicated that the average carbon stock was 0.43 kg/m2 in the organic layer, 5.30 kg/m2 at a soil depth of 0-30 cm, and 6.78 kg/m2 at a depth of 0-100 cm. The evaluation of the prediction model revealed that the average RMSE for the organic layer was 0.16 with an R2 value of 0.47, and the normalized difference vegetation index (NDVI) and forest type were identified as key environmental variables for carbon stock estimation. For the 0-30 cm soil depth, the average RMSE was 1.77 with an R2 value of 0.37, and elevation and insolation were determined to be the primary variables for the prediction model. In the prediction model for the 0-100 cm depth, the average RMSE was 2.48 with an R2 value of 0.36, and NDVI was determinedas a significant variable in addition to elevation and insolation. These findings indicate that carbon stocks in the organic layer are primarily influenced by vegetation, whereas carbon stocks in the mineral layer are predominantly governed by topographic factors such as erosion, transport, deposition, and storage of soil materials.

영어초록

This study aims to evaluate carbon stocks at varying soil depths in the forest soils of the Namyangju region using digital soil mapping techniques. The forest soils in the study area were classified into organic and mineral layers, and carbon stocks were assessed for each layer to analyze their correlation with environmental variables. The random forest algorithm was employed to predict the spatial distribution of forest soil carbon stocks. The analysis of forest soil carbon stocks in the study area indicated that the average carbon stock was 0.43 kg/m2 in the organic layer, 5.30 kg/m2 at a soil depth of 0-30 cm, and 6.78 kg/m2 at a depth of 0-100 cm. The evaluation of the prediction model revealed that the average RMSE for the organic layer was 0.16 with an R2 value of 0.47, and the normalized difference vegetation index (NDVI) and forest type were identified as key environmental variables for carbon stock estimation. For the 0-30 cm soil depth, the average RMSE was 1.77 with an R2 value of 0.37, and elevation and insolation were determined to be the primary variables for the prediction model. In the prediction model for the 0-100 cm depth, the average RMSE was 2.48 with an R2 value of 0.36, and NDVI was determinedas a significant variable in addition to elevation and insolation. These findings indicate that carbon stocks in the organic layer are primarily influenced by vegetation, whereas carbon stocks in the mineral layer are predominantly governed by topographic factors such as erosion, transport, deposition, and storage of soil materials.

참고자료

· 없음