Biomass Distribution Driving Factors And GWR Model Construction Of Pinus Massoniana Forest In Guizhou Province

Forest biomass is an important index of forest carbon sequestration capacity,and its spatial heterogeneity is affected by many factors such as natural conditions and human factors.Geographical weighted regression model(GWR)is an effective tool to study spatial variability,and it is important to explore the influencing factors of spatial heterogeneity to reveal the spatial variability of forest biomass.At the same time,analyzing the interaction and contribution rate between driving factors and forest biomass can further understand the potential influencing mechanism of forest biomass allocation.Pinus massoniana,as the main afforestation and timber species in Guizhou province,It is urgent to systematically analyze the spatial distribution characteristics and driving factors of the biomass of Pinus massoniana forest in Guizhou Province,so as to provide basic data for the biomass related research.Based on the data of the seventh continuous forest inventory in Guizhou Province,the spatial pattern of Pinus massoniana forest biomass was studied by Moran’s I,Getis-Ord Gi*analysis and Semi-variogram,and the relative contribution of biological and abiotic factors to the biomass of Pinus massoniana forest was analyzed by hierarchical partitioning analysis(AHP).Structural equation model(SEM)was used to analyze the direct and indirect effects of driving factors on the biomass of Pinus massoniana forest.The spatial modeling of driving factors and Pinus massoniana biomass was carried out using GWR model.The main conclusions are as follows:(1)Moran’s I showed that the biomass distribution of Pinus Massoniana forest in the whole province showed positive correlation in space.And it decreased with the increase of the distance.When the distance was more than 200 km,the spatial correlation of Masson pine forest biomass was very weak(P>0.05).According to local Moran’s I,high-value cluster points are located in Guiyang City,southern Zunyi City and southern Qiannan Prefecture,while low-value cluster points are mainly located in southern Qiannan Prefecture and northeastern Tongren City.The hot spot analysis shows that the hot spot is mainly located in the south of Guiyang and Zunyi,while the cold spot is mainly located in the south of Qiannan Prefecture and the northeast of Tongren.(2)The biomass of Pinus Massoniana forest had strong spatial heterogeneity,and the exponential model in the semi-variance function had the best fitting effect(R~2=0.673).The index model parameters were used as reference for the Kriging interpolation method,and the interpolation results showed that the biomass of Pinus Massoniana forest was higher in Guiyang,southern Zunyi,northeastern Anshun and southeastern Guizhou.The distribution area of the high biological value is small and scattered in patches.The biomass of Pinus Massoniana forest was mainly concentrated in southeast Guizhou and southern Guizhou,and the biomass of Masson tail pine forest in southwest Guizhou and Liupanshui region was the smallest.(3)Based on the step-up regression method,the variance expansion factor(VIF)test and Pearson correlation analysis were combined to screen out the factors that were significantly related to the biomass of Pinus Massoniana forest,including average plot diameter at breast height,stand density,Shannon-Wiener index,altitude,bedrock exposure rate,stand age,shrub cover,canopy density and stand origin.Specifically,there was a negative correlation between the exposure rate of bedrock and the biomass of Pinus Massoniana forest,while the stand age,mean DBH,stand density,canopy density,shrub coverage,altitude and Shannon index were positively correlated with the biomass of Pinus Massoniana.The biomass of Pinus Massoniana in natural forest was higher than that in plantation forest.(4)The hierarchical partitioning analysis showed that the relative contribution rate of stand density to the biomass of Pinus Massoniana was 61.13%,followed by the mean diameter at breast height of 29.5%.The stand age was 2.84%,canopy density 2.33%,Shannon-Wiener index 1.88%,stand origin 1.26%,shrub coverage 0.54%,elevation 0.28%,and bedrock exposure 0.24%.These factors could explain 72%of the biomass variation in the structural equation model.Stand density had the strongest direct effect on biomass of Pinus Massoniana,followed by mean DBH.The origin of stand,exposure rate of bedrock,canopy density and altitude had significant direct negative effects on the biomass of Pinus Massoniana forest.Shannon index was directly affected by canopy density,stand age,stand density,stand origin and shrub coverage.The shrub coverage was directly affected by the age of the stand.Canopy density was directly affected by the age of stand.Mean DBH affected Pinus Massoniana biomass indirectly through stand density,while stand age,elevation and stand origin affected Pinus Massoniana biomass indirectly through stand density.(5)The GWR model was used to construct the spatial model of Pinus Massoniana biomass.The GWR model considers the spatial autocorrelation and heterogeneity,and the results show that the sum of the squares of the GWR model is lower than the OLS model,and the fitting effect of the GWR model is better than that of the OLS model.The GWR model can effectively reduce the spatial autocorrelation of the model residuals.For the variable coefficient of horsetail pine forest biomass visualization,the results showed the strength coefficient of stand density effect are mainly distributed in north and west Zunyi and Bijie city.The mean DBH decreased from southwest to northeast of Guizhou Province.The age of the stand decreases from the center to the northwest and southeast.Shannon-Wiener index coefficient high value area is wide,mainly distributed in Zunyi city,Tongren city,Guiyang City and most areas of southeast Guizhou.The distribution of shrub coverage and canopy density coefficient was similar,and the intensity of shrub coverage decreased from southwest to northeast in Guizhou Province.The effects of elevation and basal rock exposure on the biomass of Pinus Massoniana increased from northeast to southwest.The distribution of local R~2 decreased from northwest to southeast,and the GWR model had a good fitting effect(local R~2>0.5).