Objective This study aimed to clarify the driving mechanism of carbon sink formation in natural Larix gmelinii forest, and to provide theoretical basis for improving the carbon sink function of natural L. gmelinii forest in this area.
Method Based on the survey data of 264 fixed plots from the 7th and 8th continuous national forest resource inventory in Daxing’anling region of Heilongjiang Province, northeastern China, this study selected 24 indexes from 6 aspects, namely stand, climate, diversity, soil, topography and cutting, considering 4 carbon pools of tree growth, mortality, ingrowth and cutting. The effects of each index on carbon sink in natural L. gmelinii forest were quantified by stepwise regression and structural equation model.
Result (1) The average carbon sink of natural forest of L. gmelinii from 2005 to 2010 was (1.17 ± 0.71) t/(ha·year). (2) The determination coefficient (R2), root mean square error (RMSE) and mean absolute error (MAE) of the stepwise regression model were 0.60, 0.60 t/(ha·year) and 0.45 t/(ha·year), respectively, indicating that the model was highly accurate. The variables selected by the model included Shannon-Wiener index, canopy density, soil total nitrogen, soil total potassium, average DBH of dominant trees, annual mean precipitation, slope and average age of stands. (3) In the structural equation model, the path coefficients of Shannon-Wiener index (0.462) > canopy density (0.357) > average DBH of dominant trees (0.313) > soil total nitrogen (0.286) > soil total potassium (−0.142) > annual mean precipitation (−0.107) > slope (−0.069).
Conclusion Shannon-Wiener index, canopy density and soil conditions are important driving factors affecting carbon sink. Tree species composition, average DBH of dominant trees, canopy density and soil fertility can be adjusted through rational thinning or undercrown replanting in subsequent management to improve the carbon sink function of natural forest of L. gmelinii.
DownLoad: