枫杨种源苗期高生长地理变异的趋势面分析

Trend Surface Analysis of Geographic Variations for the Seedling Height of Chinese Wing-Nut Tree ( Pterocarya stenoptera ) Provenance

采用多项式趋势面和傅里叶趋势面分析方法 ,对枫杨 15个省 (区 ) 5 5个种源 1年生苗高生长进行地理变异的模拟分析 .结果表明 ,枫杨苗高的最佳趋势面能解释 70 %以上的地理变异趋势信息量 ,并且拟合精度比传统的线性回归和线性相关分析有所提高 ;与多项式趋势面相比 ,傅里叶趋势面具有更高的拟合精度和更好的拟合效果 ;多项式趋势面一般能反映出大范围的地理变异趋势 。

The geographic variations of 55 Chinese wing nut tree provenances collected from 15 provinces or autonomous regions in China were analyzed with polynomials and Fourier Trend Analysis. The results show that geographic variations are interpreted up to 71% by the best trend surface and simulation goodness increases apparently, compared with normal linear regression and correlation models. Furthermore, the Fourier trend analysis is more accurate than the polynomials in simulation of geographic variation patterns for this flooding tolerant species. The polynomial trend analysis usually indicates macro geographic variation pattern while the Fourier trend analysis does show more detailed variations for tested provenances.