Objective  This paper aims to further understand the relationship between tree radial growth and climatic factors, reveal the relative importance as well as marginal effect of main driving climatic factors of tree radial growth, so as to enhance the application value of tree ring in the research of dendroecology and dendroclimatology.

  Method  Based on the tree ring width data of Cryptomeria japonica in Lushan Mountain area and the meteorological data of Lushan Mountain meteorological station, the tree ring width chronology was developed by the dendroclimatology method. The main driving climatic factors of tree radial growth were initially identified based on the correlation analysis method, and then the relative importance and marginal effect of the main influencing factors on Cryptomeria japonica radial growth were revealed by utilizing boosted regression tree (BRT) method.

  Result  The research results showed that the climatic factors that positively affected the radial growth of Cryptomeria japonica in Lushan Mountain, in descending order of importance, were the average minimum temperature from January to March of the current year (20.66%), the relative humidity from February to November of the previous year (15.4%), and the average vapor pressure from February to March of the current year (9.47%); the climatic factors that negatively affected the radial growth of Cryptomeria japonica in Lushan Mountain, in descending order of importance, were the sunshine hours in November of the previous year (20.81%), the maximum daily precipitation in the May of the previous year (20.54%) and the average temperature in July of the current year (13.11%). The relationship between tree ring and climate had a good linear relationship within the threshold range, and there was no linear relationship outside the threshold range.

  Conclusion  The radial growth of Cryptomeria japonica in Lushan Mountain is affected by many climatic factors, and the influence of each factor has obvious marginal effect. It is important to pay close attention to the problem of marginal effects when performing tree ring based climate reconstruction, which will enhance the reliability of tree ring climate relationship and climate reconstruction.