Hao Zang

Carbon density is an important indicator of carbon sequestration capacity in forest ecosystems. We investigated the vegetation carbon density of Pinus massoniana Lamb. forest in the Jiangxi Province. Based on plots investigation and measurement of the carbon content of the samples, the influencing factors and spatial variation of vegetation carbon density (including the tree layer, understory vegetation layer and litter layer) were analysed. The results showed that the average vegetation carbon density value of P. massoniana forest was 52 Mg·ha⁻¹. The vegetation carbon density was significantly (p < 0.01) and positively correlated with the stand age, mean annual precipitation, elevation and stand density and negatively correlated with the slope and mean annual temperature. Forest management had a significant impact on vegetation carbon density. To manage P. massoniana forest for carbon sequestration as the primary objective, near-natural forest management theory should be followed, e.g., replanting broadleaf trees. These measures would promote positive succession and improve the vegetation carbon sequestration capacity of forests. The results from the global Moran’s I showed that the vegetation carbon density of P. massoniana forest had significant positive spatial autocorrelation. The results of local Moran’s I showed that the high-high spatial clusters were mainly distributed in the southern, western and eastern parts of the province. The low-low spatial clusters were distributed in the Yushan Mountains and in the northern part of the province. The fitting results of the semivariogram models showed that the spherical model was the best fitting model for vegetation carbon density. The ratio of nugget to sill was 0.45, indicating a moderate spatial correlation of carbon density. The vegetation carbon density based on kriging spatial interpolation was mainly concentrated in the range of 32.5–69.8 Mg·ha⁻¹. The spatial distribution of vegetation carbon density regularity was generally low in the middle region and high in the peripheral region, which was consistent with the terrain characteristics of the study area.

Accurate measurement of stem diameter is essential to forest inventory. As a millimeter-level measuring tool, terrestrial laser scanning (TLS) has not yet reached millimeter-level accuracy in stem diameter measurements. The objective of this study is to develop an accurate method for deriving the stem diameter from TLS data. The methodology of stem diameter measurement by diameter tape was adopted. The stem cross-section at a given height along the stem was determined. Stem points for stem diameter retrieval were extracted according to the stem cross-section. Convex hull points of the extracted stem points were calculated in a projection plane. Then, a closed smooth curve was interpolated onto the convex hull points to simulate the path of the diameter tape, and stem diameter was calculated based on the length of the simulated path. The stems of different tree species with different properties were selected to verify the presented method. Compared with the field-measured diameter, the RMSE of the method was 0.0909 cm, which satisfies the accuracy requirement for forest inventory. This study provided a method for determining the stem cross-section and an efficient and precise curve fitting method for deriving stem diameter from TLS data. The importance of the stem cross-section and convex hull points in stem diameter retrieval was demonstrated.

Determining the response of dominant height growth to climate change is important for understanding adaption strategies. Based on 550 permanent plots from a national forest inventory and climate data across seven provinces and three climate zones, we developed a climate-sensitive dominant height growth model under a mixed-effects model framework. The mean temperature of the wettest quarter and precipitation of the wettest month were found to be statistically significant explanatory variables that markedly improved model performance. Generally, future climate change had a positive effect on stand dominant height in northern and northeastern China, but the effect showed high spatial variability linked to local climatic conditions. The range in dominant height difference between the current climate and three future BC-RCP scenarios would change from −0.61 m to 1.75 m (−6.9% to 13.5%) during the period 2041–2060 and from −1.17 m to 3.28 m (−9.1% to 41.0%) during the period 2061–2080 across provinces. The impacts of climate change on stand dominant height decreased as stand age increased. Forests in cold and warm temperate zones had a smaller decrease in dominant height, owing to climate change, compared with those in the mid temperate zone. Overall, future climate change could impact dominant height growth in northern and northeastern China. As spatial heterogeneity of climate change affects dominant height growth, locally specific mitigation measures should be considered in forest management.

The relationship of stand top and stand mean height is important for forest growth and yield modeling, but it has not been explored for natural mixed forests. Observations of stand top and stand mean height can present spatial dependence or autocorrelation, which should be considered in modeling. Simultaneous autoregressive (SAR) models, including spatial lag model (SLM), spatial Durbin model (SDM) and spatial error model (SEM), within nine spatial weight matrices were utilized to model the stand top and stand mean height relationship in the mixed Quercus mongolica Fisch. ex Ledeb. broadleaved natural stands of Northeast China, using ordinary least squares (OLS) as a benchmark model. The results showed that there was a high linear relationship between stand top and stand mean height and that there was a positive spatial autocorrelation pattern in model residuals of OLS. Moreover, SEM and SDM performed better than OLS in terms of reducing the spatial dependence of model residuals and model fitting, regardless of which spatial weight matrix was used. SEM was better than SDM. SLM scarcely reduced the spatial autocorrelation of model residuals. Among nine spatial matrices in SEM, rook contiguous matrix performed best in model fitting, followed by inverse distances raised to the second power (1/d²) and local statistics model matrix (LSM).