Ecological Statistical Study Of Pinus Tabuliformis Community In Shanxi Based On Data Transformation

Compositional data is multidimensional data that reflects the proportional structure of the data,requiring that the components of each component are not negative and the sum is a fixed value,usually 1 or 100%.The species-plot important value matrix data(hereinafter referred to as species data)conforms to the definition of compositional data and belongs to the typical compositional data.The logarithmic ratio transformation method is an effective means of mapping compositional data to Euclidean space,but it is rarely introduced in the field of quantitative ecology.Previous methods of dealing with species data were to standardize them or perform Hellinger transformation without considering the impact of the "closure effect".Topographic habitat conditions play a crucial role in the distribution of communities,and traditional classification methods mostly use only species data and ignore environmental data.In community classification,if both species and environmental data are considered,the results can reveal more deeply the relationship between plants,communities and topographic habitats.Based on the above considerations,this paper takes the Shanxi Pinus forest community as a case,then compares and discusses the impact of different processing methods of species compositional data on the results of vegetation quantity analysis by not processing species data,ALR transformation and Hellinger transformation,combined with MRT classification and CCA ranking methods,and analyzes the differences in results by exploring the ecological relationship between plant communities,species and environmental factors,and reveals the main environmental impact factors affecting Shanxi Pinus tabuliformis community.Through the above analysis,the following main conclusions are drawn:(1)Compared with the original species data,both the ALR transformation and Hellinger transformation can improve the skewness and kurtosis of the original data,reduce the probability of outliers,and make the original data tend to be normally distributed.However,the zero value in the component data is the "Achilles’ heel" of the logarithmic ratio transformation method,and the ALR transformation will be affected by the zero values.The reason is that the zero values in the species important value data are the essential zeros,especially when there are a large number of zeros in the data,which will be more tricky to handle.Hellinger transformation works well with zeros present in the data.(2)The coherence coefficients of the classification results of the three transformation methods were above 0.88,and the degree of agreement was high;the environmental factors of the divided nodes increased sequentially.Further studies were made on the plots,species divisions and environmental characteristics of the clusters,and it was found that based on the species data after the Hellinger transformation,the results of MRT classification could distinguish the differences in environmental characteristics between the plots.The quantitative analysis of species compositional data can be compared by various data transformation methods,making the division of sample communities and group naming more reasonable.(3)The influence of topographic factors and climatic factors on the community distribution of the Pinus tabuliformis community in Shanxi was more significant,and the CCA ranking results verified the results of the MRT community classification.Among them,longitude and annual hours were the determining environmental factors for the community distribution of Shanxi Pinus tabuliformis community,and the altitude,latitude,annual average temperature,annual evaporation and PH value were significantly correlated with the distribution pattern of Shanxi Pinus tabuliformis community.