Modelling Tree Recruitment In Relation To Climate In Spruce-fir Forests

The forest recruitment model is an indispensable part of the forest stand growth process.Accurately predicting the forest recruitment is the basis for accurately fitting and forecasting the forest development dynamics.However,the forest recruitment is a small probability event,with many influence factors,and it is difficult to accurately fit the forest.This study used long-term fixed sample plots of spruce-fir forests in Jingouling Forest Farm of Wangqing County.Comparing the negative binomial model,the zero-inflated model,the zero-inflated negative binomial model,the Hurdle model,and the Hurdle-negative binomial model,the sub-tree species group compares the size of the AIC value and the x2 statistic,and the Vuong test to optimize the best model.Based on traditional forest stand competition and site factors,climatic factors were added and the climate sensitive model was compared with traditional models.Through a series of evaluation indicators,the negative binomial model,the zero-inflated negative binomial model and the Hurdle-negative binomial model have certain feasibility for the model analysis of the input of the three types of tree species.Among them,the zero-inflated negative binomial model has better utility for all tree species.In all species categories,stand basal area,species basal area,and stem density have a significant effect on the occurrence of the recruitment phenomenon.In the coniferous tree species and the broad-leaved tree species group,the greater stand basal can promote the occurrence of recruitment,tree species off the area is not conducive to the occurrence of the recruitment.After adding climatic factors to the basic model,the AIC value and x2 statistic of the model have significantly decreased compared with the previous model.This shows that the model of adding climatic factors has been optimized and become more practical.In climatic factors,the annual mean temperature is positively correlated with the number of recruitment trees,and the ratio of the hottest month's temperature and the hottest month's temperature to the summer's precipitation is negatively correlated with the conifer and hardwood.Precipitation in the form of snow from August of last year to July of this year has a positive effect on the recruitment of hardwood.The zero-inflated negative binomial model have unique advantages in fitting zero-inflated data,and climatic factors cannot be ignored in the fitting process.